Дисертації з теми "EEG-TMS"

Dissertação para obtenção do Grau de Mestre em Engenharia Biomédica
Purpose: The purpose of this thesis was to provide quantitative measures of the co-registration of transcranial magnetic stimulation (TMS) and electroencephalogram (EEG). The EEG is used to study changes in the neuronal activity evoked by the non-invasive technique TMS. These effects are determined mainly based on clinical judgment. Current uses in the diagnosis of epilepsy are based only on EEG, not taking into consideration the low sensitivity in the interictal period, in particular if routine recordings are used. Methods: Patient data was gathered, analyzed and compared to healthy controls. A total of ten patients and eighteen healthy subjects underwent sessions of 75 TMS pulses. The responses to the pulses were filtered and averaged. The use of topographical scalp plots of amplitude and power, and time-series analysis of power in search for late responses provide results which enable separation of epilepsy patients and healthy controls. By investigating the significance of the results it is also possible to determine, in a quantitative way how reliable the methods are for distinguishing between the two groups. Results: The definition of what is a response is critical in this project, and as such must consider: significant power change, be above a certain amplitude, and be localized. Still, this procedure results in a non distinguishable threshold to separate both groups. Conclusions: Analysis of the receiver operating characteristic (ROC) curves also led to the understanding the method established is not entirely reliable because it cannot in fact determine differences. Since all patients were under treatment with anti-epileptic drugs(AEDs), it becomes necessary to elaborate a pilot study with recently diagnosed subjects where hyperexcitability is still present.

Prima dello sviluppo di tecniche quali quelle di Risonanza Magnetico Nucleare (RMN), di Tomografia Assiale Computerizzata (TAC) e di molteplici altre, si sono sviluppate, specifiche per l’area cerebrale, le strumentazioni di Elettroencefalografia (EEG) e di Stimolazione Magnetica Transcranica (TMS). Attraverso cinque capitoli questo elaborato fornisce una visione generale dello stato dell'arte delle strumentazioni di EEG e TMS, osservate da prima singolarmente poi accoppiate. Infine se ne osservano i possibili sviluppi futuri ed alcuni casi clinici e sperimentali correlati ad essi.

To understand how the brain functions, we must investigate the transient interactions that underpin communication between cortical regions. EEG possesses the optimal temporal resolution to capture functional connectivity, but it lacks the spatial resolution to identify the cortical locations responsible. To circumvent this problem electrophysiological connectivity should be investigated at the source level. There are many quantifiers of connectivity applied to EEG data, but some are not sensitive to the direct, or indirect, influence of one region over another, and others require the specification of a priori models so are unsuitable for exploratory analyses. Transfer Entropy (TE) can be used to infer the direction of linear and non-linear information exchange between signals over a range of time-delays within EEG data. This thesis explores the creation of a new method of mapping dynamic brain connectivity using a trial-based TE analysis of EEG source data, and the application of this technique to the investigation of semantic and number processing within the brain. The first paper (Chapter 2) documents the analyses of a semantic category and number magnitude judgement task using traditional ERP techniques. As predicted, the well-known semantic N400 component was found, and localised to left ATL and inferior frontal cortex. An N365 component related to number magnitude judgement was localised to right superior parietal regions including the IPS. These results offer support for the hub-and-spoke model of semantics, and the triple parietal model of number processing. The second paper (Chapter 3) documents an analysis of the same data with the new trial-based TE analysis. Word and number data were analysed at 0-200ms, 200-400ms, and 400-600ms following stimulus presentation. In the earliest window, information exchange was occurring predominately between occipital sources, but by the latest window it had become spread out across the brain. Task-dependent differences of regional information exchange revealed that temporal sources were sending more information to occipital sources following words at 0-200ms. Furthermore, the direction and timing of information movement within a front-temporal-parietal network was identified during 0-400ms of the number magnitude judgment. The final paper (Chapter 4), documents an attempt to track the influence of TMS through the brain using the TE analysis. TMS was applied to bilateral ATL and IPS because they are both important hubs in the brain networks that support semantic and number processing respectively. Left ATL TMS influenced sources located primarily in wide-spread left temporal lobe, and inferior frontal and inferior occipital cortices. The anatomical connectivity profile of the temporal lobe suggests that these are all plausible locations, and they exhibited excellent spatial similarities to the results of neuroimaging experiments that probed semantic knowledge. The analysis of right ATL TMS obtained a mirror image of the left. Left parietal stimulation resulted in a bilateral parietal, superior occipital, and superior prefrontal influence, which extended slightly further in the ipsilateral hemisphere to stimulation site. A result made possible by the short association and callosal fibres that connect these areas. Again, the results at the contralateral site were a virtual mirror image. The thesis concludes with a review of the experimental findings, and a discussion of methodological issues still to be resolved, ideas for extensions to the method, and the broader implications of the method on connectivity research.

Affordances play a part in how we prepare to handle objects. Tools and other manipulable objects are said to automatically “afford” various actions depending upon the motor repertoire of the actor. Evidence obtained through behavioural experiments, fMRI, EEG and TMS has proven that this is the case but, as yet, the temporal evolution of affordances has not been fully investigated. Determining the critical time-scale may have significance to patients with brain damage or motor disorders when attempting object manipulation. There are many other factors involved in therapy but it is worth considering that there could be an optimum period of time to view an object before the benefit of an automatic affordance is no longer available. In a series of experiments using the novel approach of positioning the participant’s dominant hand closer to or further from the object being viewed, together with use of three dimensional stimuli, and through application of behavioural assays, TMS pulses and EEG recordings, this research examined temporal properties of affordances in young healthy control subjects. Verification of this motoric activity by EEG led to investigating chronic phase stroke survivors with remaining upper limb deficits and comparing their brain activity with age-matched control participants. As EEG and TMS both have good temporal properties, they are ideal converging methodologies for this kind of investigation. By mapping how affordances develop and dissipate, this work has yielded pure scientific advances in the field of motor decision making. Further, it has resulted in suggestions for future research relating to a possible method to improve rehabilitation interventions for patients who are neurologically impaired by stroke.

Approximately 30% of epilepsy patients do not respond to treatment, whereas others attain seizure freedom with the same medication, the reasons remain unclear. Previous work utilised transcranial magnetic stimulation (TMS) to reveal predictive markers of treatment response in new-onset drug-naïve patients, distinguishing good responders from those who become intractable. I hypothesised that these markers should also be present in long-term treatment resistant epilepsy, allowing outcome prediction in patients commencing new medications. A central hypothesis in this thesis is that, interictal brain dynamics in epilepsy differ from the stable state of the healthy brain and are related to seizure frequency. I addressed this using TMS measures of excitation and inhibition, and electroencephalography (EEG) as a measure of the larger-scale electrophysiological dynamic system. Using existing TMS data I examined motor evoked potentials (MEPs) in Idiopathic generalised epilepsy (IGE). MEPs were more polyphasic in patients and their relatives than controls, which may represent an inherited endophenotype. TMS measures were also compared between patients with well and poorly controlled epilepsy. Findings broadly indicated that poorly controlled patients have reduced excitability vs well controlled, the reasons are unknown, although a protective homeostatic mechanism may be responsible. Furthermore, TMS parameters in well-controlled epilepsy were closer to healthy controls than poorly controlled patients. A longitudinal TMS study in chronic epilepsy with patients studied before and after treatment change, revealed a weak effect suggesting reduced excitability in poor responders, although a range of factors suggested that TMS would not have utility as a predictive marker clinically. A pilot study also investigated whether external trigeminal nerve stimulation (eTNS) has a measureable effect on brain excitability. There was a small effect which may associate with treatment outcome. Finally, EEG measures were compared in well and poorly-controlled epilepsy, with the profile of a well-controlled patient closer to that of the healthy brain. Future work focusing on EEG as a marker of response in newly diagnosed epilepsy, utilising TMS-EEG for revealing mechanisms underlying treatment response would be appropriate.

Transkranijinė magnetinė stimuliacija (TMS) – tai modernus neinvazinis vaistams rezistentiškų psichiatrinių sutrikimų gydymo būdas. Fiziologiniai TMS tyrimai pasižymi įvairiais, dažnai prieštaringais rezultatais, daugeliu atvejų didžiausias dėmesys skiriamas betarpiškiems poveikiams po vienos TMS procedūros, bet ne po pilno terapinio kurso. Manoma, kad rezultatų įvairovę TMS praktikoje įtakoja skirtingi stimuliacijos parametrai ir netikslumai parenkant stimuliuojamą zoną smegenyse. Nors TMS terapija dažnai traktuojama kaip švelnesnė alternatyva elektros impulsų terapijai (EIT), palyginamųjų fiziologinių šių metodikų tyrimų labai trūksta. Darbo tikslas buvo įvertinti TMS terapijos kurso poveikį bioelektriniam galvos smegenų aktyvumui ir palyginti jį su EIT terapijos poveikiu. Buvo tirta aukšto ir žemo dažnių (10 Hz ir 1 Hz) TMS terapijos įtaka EEG dažnių galios spektrui bei sukeltiniam klausos potencialui P300, naudojant standartinį ir neuronavigacinį taikinio pozicionavimą. TMS sukelti EEG pokyčiai palyginti su EIT terapijos sukeltais EEG pokyčiais, išmatuota TMS terapijos sąlygotų pokyčių dinamika kelių mėnesių bėgyje. Rezultatai parodė, kad TMS terapijos pasekoje smegenyse ryškiausiai padidėja delta dažnio galia. Naudojant standartinį pozicionavimą 10 Hz TMS sukėlė įvairesnius ir intensyvesnius EEG galios spektro pokyčius nei 1 Hz TMS. Pritaikius neuronavigacinę sistemą 10 Hz TMS atveju sumažėjo teta ir alfa dažnių galios pokyčiai. Praėjus keliems mėnesiams nuo TMS... [toliau žr. visą tekstą]
Transcranial magnetic stimulation (TMS) is a modern non invasive method of drug resistant psychiatric disorder treatment. TMS physiology research is hindered by variable, often controversial results. In most studies main attention is being focused on immediate effects after single TMS procedure rather than the influence of a complete therapy course. It is considered that variability of results in TMS practice is caused by different stimulation parameters and imprecision of stimulated area placement in the brain. Although TMS therapy is often viewed as a milder alternative to electroconvulsive therapy (ECT), comparative physiological studies of these two methods are very rare. The aim of this study was to evaluate the effect of rTMS therapy course on bioelectrical brain activity and compare it to an ECT effect. Research included the effect of high and low frequency (10 Hz and 1 Hz) TMS on EEG band power spectrum and auditory evoked potential P300, using both standard and neuronavigated target positioning. TMS evoked EEG changes were also compared to the changes of ECT. Change dynamics after several months of TMS therapy were also measured. Results showed that after TMS therapy the most notable change in the brain occurs in the form of delta power increase. When using standard positioning 10 Hz TMS evokes more diverse and intense EEG band power spectrum changes than the 1 Hz TMS. Application of neuronavigation system decreases theta and alpha band power changes in 10 Hz TMS... [to full text]

Transcranial magnetic stimulation (TMS) is a modern non invasive method of drug resistant psychiatric disorder treatment. TMS physiology research is hindered by variable, often controversial results. In most studies main attention is being focused on immediate effects after single TMS procedure rather than the influence of a complete therapy course. It is considered that variability of results in TMS practice is caused by different stimulation parameters and imprecision of stimulated area placement in the brain. Although TMS therapy is often viewed as a milder alternative to electroconvulsive therapy (ECT), comparative physiological studies of these two methods are very rare. The aim of this study was to evaluate the effect of rTMS therapy course on bioelectrical brain activity and compare it to an ECT effect. Research included the effect of high and low frequency (10 Hz and 1 Hz) TMS on EEG band power spectrum and auditory evoked potential P300, using both standard and neuronavigated target positioning. TMS evoked EEG changes were also compared to the changes of ECT. Change dynamics after several months of TMS therapy were also measured. Results showed that after TMS therapy the most notable change in the brain occurs in the form of delta power increase. When using standard positioning 10 Hz TMS evokes more diverse and intense EEG band power spectrum changes than the 1 Hz TMS. Application of neuronavigation system decreases theta and alpha band power changes in 10 Hz TMS... [to full text]
Transkranijinė magnetinė stimuliacija (TMS) – tai modernus neinvazinis vaistams rezistentiškų psichiatrinių sutrikimų gydymo būdas. Fiziologiniai TMS tyrimai pasižymi įvairiais, dažnai prieštaringais rezultatais, daugeliu atvejų didžiausias dėmesys skiriamas betarpiškiems poveikiams po vienos TMS procedūros, bet ne po pilno terapinio kurso. Manoma, kad rezultatų įvairovę TMS praktikoje įtakoja skirtingi stimuliacijos parametrai ir netikslumai parenkant stimuliuojamą zoną smegenyse. Nors TMS terapija dažnai traktuojama kaip švelnesnė alternatyva elektros impulsų terapijai (EIT), palyginamųjų fiziologinių šių metodikų tyrimų labai trūksta. Darbo tikslas buvo įvertinti TMS terapijos kurso poveikį bioelektriniam galvos smegenų aktyvumui ir palyginti jį su EIT terapijos poveikiu. Buvo tirta aukšto ir žemo dažnių (10 Hz ir 1 Hz) TMS terapijos įtaka EEG dažnių galios spektrui bei sukeltiniam klausos potencialui P300, naudojant standartinį ir neuronavigacinį taikinio pozicionavimą. TMS sukelti EEG pokyčiai palyginti su EIT terapijos sukeltais EEG pokyčiais, išmatuota TMS terapijos sąlygotų pokyčių dinamika kelių mėnesių bėgyje. Rezultatai parodė, kad TMS terapijos pasekoje smegenyse ryškiausiai padidėja delta dažnio galia. Naudojant standartinį pozicionavimą 10 Hz TMS sukėlė įvairesnius ir intensyvesnius EEG galios spektro pokyčius nei 1 Hz TMS. Pritaikius neuronavigacinę sistemą 10 Hz TMS atveju sumažėjo teta ir alfa dažnių galios pokyčiai. Praėjus keliems mėnesiams nuo TMS... [toliau žr. visą tekstą]

La Stimolazione transcranica a Corrente Diretta continua (tDCS) è una tecnica di neurostimolazione non invasiva in grado di generare alterazioni dell’eccitabilità neuronale dipendenti dalla plasticità. Negli ultimi anni si sta assistendo a un crescente interesse nell’utilizzo di questa tecnica, sia in settings clinici che sperimentali. In particolare, la possibilità di indurre effetti a lungo termine rende la tDCS interessante nel trattamento di deficit cognitivi associati a disturbi sia neuropsicologici che psichiatrici. Nonostante la sua crescente diffusione, si sa ancora poco riguardo i meccanismi neurofisiologici alla base del suo funzionamento, soprattutto per quanto riguarda gli effetti su regioni cerebrali che sottostanno a funzioni cognitive di ordine superiore. Una conoscenza più approfondita dei meccanismi alla base della tDCS potrebbe quindi essere cruciale per riuscire a definire e migliorare i protocolli di stimolazione sia clinici che di ricerca. A questo scopo, uno studio sistematico degli effetti corticali della tDCS appare ancora più necessario. In questo progetto abbiamo quindi esplorato gli effetti di plasticità corticale indotti dalla stimolazione catodica in soggetti sani, sia a riposo che durante l’esecuzione di un compito, utilizzando il sistema integrato di Stimolazione Magnetica Transcranica ed Elettroencefalografia (TMS-EEG). Il TMS-EEG è una tecnica molto efficace in quando permette di misurare direttamente la modulazione dell’eccitabilità corticale su tutta la corteccia. In questa tesi sono riportati tre studi. Nel primo, partendo dai risultati sulla tDCS anodica già precedentemente pubblicati, sono stati esplorati gli effetti della tDCS catodica sulla Corteccia Parietale Posteriore (PPC) destra a riposo. Nel secondo studio, abbiamo indagato gli effetti comportamentali indotti dalla tDCS catodica sulla PPC destra, durante l’esecuzione di due compiti: uno di memoria di lavoro visuospaziale e uno di orientamento dell’attenzione visuospaziale. Nel terso studio, infine, abbiamo utilizzato nuovamente il TMS-EEG per tracciare gli effetti neurofisiologici della tDCS catodica sulla PPC di destra mentre i partecipanti erano impegnati nell’esecuzione dei compiti individuati nello studio precedente. I risultati per la tDCS catodica a riposo non hanno mostrato effetti di modulazione dell’eccitabilità corticale, sia a livello dei sensori che a livello delle sorgenti, sia a livello locale che globale. I risultati precedentemente ottenuti con la tDCS anodica, hanno invece mostrato un aumento diffuso dell’eccitabilità corticale lungo un network bilaterale frontoparietale, che rispecchia le connessioni strutturali tra le aree d’interesse. Durante l’esecuzione di un compito, invece, la stimolazione catodica, così come quella anodica, ha mostrato una modulazione nell’eccitabilità corticale solo in quelle aree che sono coinvolte nell’esecuzione del compito. In conclusione, in questo lavoro di tesi emergono diversi interessanti risultati. Innanzitutto, questi dati mostrano un effetto non lineare della tDCS sull’eccitabilità corticale a riposo, che non possono essere completamente spiegati dal semplice dualismo anodico-eccitatorio catodico-inibitorio. Un altro risultato rilevante è dato dal ruolo cruciale giocato dai differenti stati ti attivazione corticale (a riposo Vs attivo). Questi risultati sembrano mostrare che il livello di attivazione corticale di base contribuisca a modulare gli effetti della stimolazione, in accordo con l’ipotesi “attività-selettività”. Lo stato di attivazione di base deve quindi essere preso in considerazione, in particolare se si vogliono osservare degli effetti di neuromodulazione con tDCS catodica. In generale, tutte queste osservazioni contribuiscono a costruire quel corpus di conoscenze necessario soprattutto per la definizione dei parametri tDCS sia per esperimenti di neuroscienze cognitive che per protocolli riabilitativi.
Transcranial direct current stimulation (tDCS) is a non-invasive neuromodulatory technique able to induce plasticity-related alterations in neuronal excitability. There is a growing interest in the use of tDCS in both experimental and clinical settings; in particular, the chance to induce long-term effects fostered the used of the technique to treat cognitive impairments associated with different neuropsychological and psychiatric disorders. Although tDCS is increasingly used, presently little is still known about its neurophysiological underpinnings, particularly concerning the activity on the brain regions that underlie high cognitive brain functions. In these cases, optimal tDCS stimulation parameters also have yet to be clearly defined. A deeper understanding of the mechanisms underpinning this technique would be crucial to achieving a better refinement of stimulation protocols for clinical and research purposes. For this reason, a systematic and comprehensive study of its cortical effects acquires a critical relevance. In the last years, there has been indeed a keen interest in understanding the working mechanisms of this technique. To address this issue, in this project we explored the cortical plasticity modulation induced by cathodal stimulation on healthy subjects while resting or during task execution, using an integrated system of Transcranial Magnetic Stimulation and Electroencephalography (TMS-EEG), which allows to directly measure cortical excitability modulation all over the cortex and effective connectivity. In the first study, starting from the previous results with anodal tDCS, the effects of cathodal stimulation over the right Posterior Parietal Cortex (PPC) were explored during resting state. The contralateral homologue brain area, namely the left posterior parietal cortex (PPC), was targeted with TMS before, during, and after cathodal stimulation. In the second study, we explored the behavioural effects induced by the application of cathodal tDCS over right PPC during the execution of two tasks, one of visuospatial working memory and a second tapping visual attention reorienting, which are known to involve this brain area. The aim was to find tasks sensitive to the effect of cathodal tDCS over the right PCC, to be used in the third study. A disruption of the performance was found for the Posner Cueing Task. In the third study, we employed again TMS-EEG to track the neurophysiological effects of cathodal tDCS on right PPC at an active state, i.e. while the participants were performing the task tested on the second study. The results at resting state for cathodal tDCS, both at sensors and cortical sources levels, converge in showing no differences during and after tDCS compared to pre-stimulation sessions, both at a global and local level. The previous results with anodal tDCS, instead, reported a widespread rise of cortical excitability along with a bilateral frontoparietal network, following structural connections. On the other hand, at an active state, cathodal, as well as anodal, tDCS induced modulation of cortical excitability only in the task-relevant brain regions. Several significant findings emerged from this empirical work. First of all, these data highlight a non-linear impact of anodal and cathodal stimulation on cortical excitability at rest that is not depicted by the simplistic view of anodal-excitatory and cathodal-inhibitory effects. Another relevant point is the crucial role played by the different cortical states (resting vs active). These results seem to point out that the level of cortical state can contribute to modulate the tDCS effects, in line with “activity-selectivity” hypothesis. The level of cortical state needs to be taken into account, especially to observe neuromodulatory effects also with cathodal tDCS. All these findings hold relevant implications for tDCS setup in both cognitive neuroscience experiments and rehabilitation protocols.

Le oscillazioni neurali sono considerate elementi costitutivi del funzionamento cognitivo, e negli ultimi decenni i neuroscienziati hanno sviluppato teorie fondamentali sul ruolo delle oscillazioni nelle dinamiche cerebrali. Recentemente, un crescente numero di evidenze ha mostrato come l’attività oscillatoria in corso possa rendere conto di una porzione considerevole della variabilità che si osserva nella prestazione comportamentale e nella risposta neurofisiologica. Nel dominio della percezione visiva, un ruolo cruciale è svolto dalle oscillazioni neurali nel range di frequenza alfa. Si ritiene che l’attività alfa eserciti una funzione inibitoria sull’elaborazione dello stimolo e rifletta l’eccitabilità corticale. E’ stato recentemente proposto che il ritmo alfa non possa essere considerato come un fenomeno unitario; tuttavia, si conosce ancora poco riguardo ai meccanismi neurali associati con l’attività alfa misurata attraverso registrazioni non invasive. Inoltre, fino ad ora la maggior parte degli studi sugli effetti dell’attività in corso sulla percezione visiva si è focalizzata su una particolare classe di stimoli, ovvero che hanno un’intensità vicino alla soglia sensoriale, e si conosce molto meno riguardo a cosa avvenga in risposta a stimoli la cui intensità va oltre il valore di soglia. Nel presente lavoro, ci siamo posti l’obiettivo di affrontare tali questioni studiando gli effetti dell’attività alfa in corso sulla risposta percettiva e neurofisiologica nel dominio visivo. Il primo obiettivo era quello di replicare alcune evidenze sugli effetti del power e della fase delle fluttuazioni spontanee dell’attività alfa pre-stimolo sulla detezione visiva, utilizzando stimoli a soglia. In aggiunta allo studio originale, l’utilizzo della magnetoencefalografia ci ha permesso di ricostruire le sorgenti cerebrali dell’attività oscillatoria pre-stimolo e dell’attività evocata. Un secondo studio era volto a modulare l’attività alfa in corso utilizzando un paradigma di deprivazione sensoriale, e testare gli effetti di tale modulazione attraverso un ampio range di intensità di stimolazione. L’uso della stimolazione magnetica transcranica (TMS) con simultanea registrazione elettroencefalografica ci ha permesso di valutare la risposta neurofisiologica e percettiva alla TMS, attraverso i potenziali evocati e la percezione dei fosfeni. Infine, in un terzo studio abbiamo sviluppato un modello formale sugli effetti dell’attività alfa in corso sulla percezione visiva, con l’obiettivo di distinguere possibili meccanismi neurali che non possono essere disambiguati a livello non-invasivo. Il modello è basato sulle interazioni cross-frequency tra l’inibizione funzionale di alfa e l’attività gamma dei neuroni sensoriali, e mette in evidenza i vantaggi di presentare un ampio range di intensità di stimoli nello studio degli effetti dell’attività oscillatoria, utilizzando un approccio psicofisico. Considerati insieme, i nostri risultati sono coerenti con la letteratura corrente riguardo alla funzione inibitoria svolta dall’attività alfa in corso sulla percezione visiva. Infatti, la risposta sia percettiva che neurofisiologica ad uno stimolo esterno era influenzata dall’attività alfa pre-stimolo, nelle fluttuazioni spontanee così come quando era modulata da un paradigma di deprivazione sensoriale. Inoltre, le presenti evidenze supportano l’ipotesi che le oscillazioni alfa sottendano meccanismi distinti, e mettono in luce come nuove intuizioni possano emergere dall’utilizzo di un approccio psicofisico allo studio dell’attività oscillatoria in corso sulla percezione. Utilizzando diversi approcci metodologici, il presente lavoro fornisce nuovi avanzamenti nello studio non-invasivo delle oscillazioni sul comportamento, nello specifico sull’inibizione dell’attività alfa sulla percezione visiva.
Neural oscillations are considered to be the building blocks of cognitive functioning, and in the last decades neuroscientists have developed fundamental theories on their role in brain dynamics. Recently, a growing body of evidences has shown that ongoing oscillatory activity can account for a considerable amount of variability in behavioral performance and in neurophysiological response. In the domain of visual perception, a crucial role is played by neural oscillations within alpha frequency range. Alpha activity is believed to exert an inhibitory function on stimulus processing and to reflect cortical excitability, both when it fluctuates spontaneously as well as when it is modulated, by top-down or bottom-up mechanisms. It has been recently suggested that alpha rhythm may not be considered as a unitary phenomenon; however, still little is known about the neural mechanisms associated with alpha activity as measured by non-invasive recordings. Furthermore, up to now most of the studies on the effects of ongoing alpha activity on visual perception focused on a special class of stimuli, i.e., with a near-threshold intensity, and much less is known about what happens in the response beyond sensory threshold. In the present work, we aimed at addressing these issues by studying the effects of ongoing alpha oscillations on perceptual and neurophysiological outcome in the visual domain. The first goal was to replicate recent findings on the effects of spontaneous fluctuations of pre-stimulus alpha power and phase on a visual detection task, by using near-threshold stimuli. In addition to the original study, the use of magnetoencephalography allowed us to reconstruct brain sources of pre-stimulus and evoked activity. In a second study, we aimed at modulating ongoing alpha activity by using a sensory deprivation paradigm, and tested the effects of such modulation by means of a wide range of stimulation intensities. The use of transcranial magnetic stimulation (TMS) with concurrent electroencephalography allowed to directly assess the neurophysiological and perceptual response to TMS, by means of TMS-evoked potentials and phosphene perception. Finally, in a third study we developed a formal model of the effects of ongoing alpha activity on visual perception, with the aim of disentangling possible neural mechanisms which cannot be discerned non-invasively. The model is based on cross-frequency interactions between alpha functional inhibition and gamma activity of sensory neurons and highlights the advantages of presenting a wide range of stimulus intensities in the study of the effects of pre-stimulus oscillatory activity, using a psychophysical approach. Taken together, our results are consistent with current literature about the inhibitory function played by ongoing alpha activity on visual perception. Indeed, both perceptual and neurophysiological response to an external stimulus were affected by pre-stimulus alpha activity, when it fluctuated spontaneously as well as when it was modulated by a sensory deprivation paradigm. Moreover, the present findings support the hypothesis that alpha oscillations subtend distinct mechanisms, and highlighted that new insights may arise from applying a psychophysical approach to the study of ongoing activity on perception. By using different methodological approaches, the present work provides novel advances in the field of non-invasive investigation of ongoing oscillations on behavior, specifically on alpha inhibition of visual perception.

The limited evidence and inconsistency of purposeful behaviors in patients in a minimally conscious state (MCS) asks for objective electrophysiological marker of the level of consciousness. Here, a comparison between event-related potentials (ERPs) was investigated using different level of stimulus complexity. ERPs were recorded in seventeen patients, 6 of which in vegetative state (VS), 11 in MCS, and 10 controls. Three oddball paradigms with different level of complexity were applied: sine tones, the subject’s own name versus sine tones and other first names. Latencies and amplitudes of N1 and P3 waves were compared. Cortical responses were found in all MCS patients, and in 6 of 11 patients in VS. Healthy controls and MCS patients showed a progressive increase of P3 latency in relation to the level of stimulus complexity. No modulation of P3 latency was observed in the vegetative patients. These results suggest that the modulation of P3 latency related to stimulus complexity may represent an objective index of higher-order processing integration that predicts the recovery of consciousness from VS to MCS when clinical manifestations are inconsistent. A second step was encouraged by the work of Schiff et al. (2007) reporting a MCS patient who responded to deep brain stimulation (DBS). We explored six patients that participated in an ABA design alternating between repetitive transcranial magnetic stimulation (rTMS) and peripheral nerve stimulation. After peripheral stimulation, patients did not exhibit clinical, behavioral, or electroencephalographic (EEG) changes. The frequency of specific and meaningful behaviors increased after rTMS in a patient, along with the absolute and relative power of the EEG δ, β, and α bands. Afterwards, a more consistent sample has been enrolled to reproduce the first encouraging results. Thirty MCS/VS patients participated to a randomized controlled trial consisting of transcranial stimulations with transcranial direct current stimulation (tDCS) and rTMS. Patients in MCS showed an increase of long range fronto-parietal connectivity indicating a complex information processing and a decrease of fluctuation of arousal . VS patients did not. These results suggest that rTMS may improve long range connections between remote cortical areas and promote, at some level, recovery of awareness and arousal in MCS patients.
Le limitate evidenze e la fluttuazione dei comportamenti intenzionali neiin pazienti in stato di minima coscienza (SMC) richiedono la ricerca di un indice marcatore elettrofisiologico obiettivo del livello di coscienza. Nel presente studio, è stato mostrato un confronto tra potenziali evento-correlati (ERP) utilizzando diversi livelli di complessità di stimolo. Gli ERP sono stati registrati in diciassette pazienti, di cui sei in stato vegetativo (SV), 11 in SMC, e 10 controlli sani. I partecipanti sono stati sottoposti a tre paradigmi di diverso grado di complessità: toni puri, il nome proprio del soggetto verso toni puri, e verso altri nomi. Sono state riscontrate risposte corticali in tutti i pazienti in SMC, ed in 6 degli 11 pazienti in SV. I controlli sani ed i pazienti in SMC hanno mostrato un progressivo aumento della latenza dell’onda P300 in relazione al livello di complessità dello stimolo. Nessuna modulazione di latenza è stata osservata nei pazienti in SV. Questi risultati suggeriscono che la modulazione di latenza della P300 relativa a complessità dello stimolo può rappresentare un indice obiettivo dell’integrazione tra aree di elaborazione di ordine superiore, presupposto necessario per il recupero della coscienza. Un secondo passo è stato incoraggiato dal lavoro di Schiff e coll. (2007) che riportarono il miglioramento clinico di un paziente in SMC dopo stimolazione cerebrale profonda (DBS). Abbiamo studiato sei pazienti sottoponendoli ad uno studio di tipo ABA con alternanza tra stimolazione magnetica transcranica ripetitiva (rTMS) e stimolazione dei nervi periferici. Dopo stimolazione periferica, i pazienti non ha evidenziato variazioni dei quadric clinico, comportamentale o elettroencefalografico (EEG). Tuttavia, dopo la rTMS, un paziente manifestò un aumento della frequenza di specifici comportamenti coscienti, associato ad un incremento della potenza assoluta e relativa delle bande EEG alfa, beta e delta. Successivamente, è stato arruolato un campione più consistente di pazienti per riprodurre i primi incoraggianti risultati. Trenta pazienti in SV/SMC hanno partecipato ad uno studio controllato randomizzato che comportava l’utilizzo di stimoli transcranici con stimolazione transcranica a corrente continua (tDCS) e rTMS. I pazienti in SMC hanno mostrato un aumento di connettività fronto-parietale, che indica una complessa elaborazione delle informazioni sensoriali, ed una diminuzione della fluttuazione dell’arousal. Il quadro dei pazienti in SV rimase invariato. Questi risultati suggeriscono che la rTMS può migliorare le connessioni a lungo raggio tra remote aree corticali e promuovere, in qualche modo, il recupero di coscienza nei pazienti in SMC.

ABSTRACT In the context of increasing incidence of stroke (but also an increasing rate of survival), non-invasive brain stimulation techniques (NIBS) are more frequently used for patients with post-stroke aphasia (PWA) and post-stroke depression (PSD). NIBS techniques, modulating brain plasticity, might offer valid, alternative therapeutic strategies. The aim is to reach a better outcome because treatment of aphasia can also improve post-stroke depression and vice versa. Based on two literature reviews on NIBS effects on PSD and post-stroke aphasia the conclusion is that, although the field is relatively new, and many more investigations with larger samples of patients are required, transcranial magnetic stimulation (TMS) and transcranial direct current stimulation (tDCS) clinical application is well tolerated, safe, and feasible. Starting from these encouraging data, we used a combination of TMS and electroencephalography (EEG) to explore the excitability modulation before and after active (20 sessions) and sham (20 sessions) tDCS in a double-blind crossover experiment. Four chronic non fluent PWA underwent 8 weeks of verbal exercises coupled with tDCS over the perilesional areas close to the left inferior frontal gyrus. To evaluate changes induced by tDCS, TMS-EEG responses over Brodmann area 6 (BA6) were computed using five different parameters. In addition, these data were compared with those recorded from a matched control group. The results indicated a slight improvement after tDCS stimulation (as compared to sham) for patients with Broca’s aphasia, but not for those with global aphasia. Also, TMS-evoked EEG responses recorded from the ipsilesional hemisphere were abnormal in individuals with chronic post-stroke aphasia (slower and simple responses with higher amplitudes) when compared to responses from the contralesional hemisphere and from the control group. Critically, the Global Mean Field Power (GMFP), Local Mean Field Power (LMFP) and Natural Frequency values were modulated by anodal tDCS. Despite these interesting results, further data are needed in order the obtain more direct, stronger evidence linking behavioral tDCS effects and neurophysiological data.

It is well established that ageing is associated with a decline in manual dexterity. An important neural process for the control of manual dexterity is motor cortical inhibition, which is the process by which neural activity within the motor cortex is supressed. Reductions in motor cortical inhibition may contribute to the age-related decline in manual dexterity. Paired-pulse transcranial magnetic stimulation (TMS) can be used to measure long-interval intracortical inhibition (LICI) in the motor cortex. Previous literature examining differences in LICI between young and older adults have produced conflicting results. In addition, none have included a middle-aged group of participants. The purpose of the current study was to determine whether there are differences in LICI between young and middle-aged adults. An emerging technique that combines TMS with electroencephalography (EEG) was used to measure LICI. In 12 young and 13 middle-aged participants, the TMS-evoked potential (TEP; recorded from EEG) reflected the motor cortical response to sham TMS, single-pulse TMS, and paired-pulse TMS. The TEPs generated by single- and paired-pulse TMS did not differ between young and middle-aged adults. Therefore, there is no evidence from the current study to suggest differences in motor cortical inhibition between young and middle-aged adults. However, these results are speculative as the TEPs generated by sham and single-pulse TMS were highly similar, suggesting that artefacts heavily influenced the TEPs. It is critical that future studies are able to minimise the artefacts during TMS-EEG recording, and reliably identify and remove artefacts from the EEG data.

La dextérité, notamment la pince de précision (i.e., opposition pouce-index) est une fonction très développée chez l’homme. Elle est basée sur l’habileté à contrôler précisément et indépendamment les forces et mouvements des doigts en relation avec les contraintes de la tâche. Les muscles de la main responsables du mouvement des doigts sont gouvernés par le système corticospinal (CS) latéral. La principale source de ce système CS est l’aire motrice primaire (M1), laquelle possède des projections CS directes sur les motoneurones des muscles de la main. Cependant, d’autres projections CS en provenance des aires motrices non primaires ont été trouvées, notamment en provenance de l’aire motrice supplémentaire (SMA). Chez l’homme, la fonctionnalité de cette voie dans le contrôle habile des doigts a peu été étudiée. L’objectif de cette thèse est d’étudier, chez l’homme, l’implication des projections CS de la SMA lors de contrôle manuel précis de force. Pour ce faire, nous avons utilisé la stimulation magnétique transcrânienne (TMS) et l’électroencéphalographie (EEG).A travers différentes études, nous avons pu mettre en évidence l’importante implication de la SMA dans la dextérité. Il semblerait que cette aire puisse agir en parallèle à M1 en régulant directement l’excitabilité des motoneurones de la moelle épinière. En conclusion, nos résultats suggèrent que M1 et SMAp ont une influence directe et efficace sur la production de force pendant des tâches motrices manuelles fines
Human dexterity is a highly developed function based on the ability to independently and precisely control forces and movements of the fingers related to the constraints of the task. Hand muscles for finger movements are steered by the lateral corticospinal (CS) system. The main source of this CS system is the primary motor area (M1), which has direct CS projections on motoneurons innervating hand muscles. Recently, CS projections from non-primary motor area have also been found, especially from the supplementary motor area (SMA). However, the functionality of this CS tract in human manual force control is little studied. The aim of this thesis was to study the implication of the CS projections from SMA in precision manual force control, using electroencephalography (EEG) and transcranial magnetic stimulation (TMS).Altogether, the results obtained in our different studies show the important implication of SMA in dexterity. It appears that this area can act in parallel with M1, directly influencing excitability of spinal motoneurons. We conclude that M1 and SMA both have direct and efficient influence on force production during fine manual motor tasks

La sinestesia è un raro fenomeno percettivo che consiste nell’interazione e sovrapposizione spontanea di più sensi: la stimolazione di una modalità sensoriale induce automaticamente una percezione in una seconda modalità, anche in assenza di una reale stimolazione di quest’ultima. La tesi tratterà inizialmente del fenomeno della sinestesia nel suo complesso e dei modelli neurali proposti al fine di comprendere le basi neurali sottostanti ad essa, dopodichè verranno approfondite alcune delle forme di sinestesia esistenti e verranno poi affrontate le tecniche di indagine cerebrale impiegate nello studio della sinestesia. Infine verranno illustrate tre nuove tecnologie basate sui principi della sinestesia, utilizzate in ambito terapeutico.

Dans la vie de tous les jours, il arrive que nos actions soient perturbées par desvariations rapides des forces externes de notre environnement. Afin d'atteindre notre but, nousdevons alors réagir “vite et bien” à ces perturbations de mouvement, ce qui implique la mise enjeu à la fois de processus cognitifs et de processus sensori-moteurs. Nous nous sommesintéressés aux mécanismes corticaux (engagés notamment au niveau du cortex sensorimoteurprimaire) sous-tendant les interactions entre fonctions cognitive et sensorimotrice permettantd'adapter la réaction à la perturbation en fonction de notre intention, en nous efforçant de fairele lien entre les mécanismes impliqués au cours de la préparation et de la réalisation de laréaction. En utilisant le couplage EEG-TMS (avec enregistrement de l'EMG), nous avons menéune approche par stimulation-enregistrement, permettant d'observer simultanément lesmécanismes corticaux et corticospinaux précédant et suivant la stimulation, et ainsi de mieuxcomprendre le lien reliant l'activité cérébrale et le comportement.Dans l’étude 1, nous avons utilisé une perturbation motrice centrale, c'est-à-dire quenous avons demandé au sujet soit de résister soit d'assister un mouvement évoqué directementau niveau cortical par TMS. Ceci nous a permis de montrer que les processus cognitifs peuventinfluencer directement l'excitabilité corticale et corticospinale, avant la mise en jeu deprocessus sensorimoteurs impliqués dans l’exécution du mouvement. Lorsque le sujet s’estpréparé à résister au mouvement évoqué par TMS, l'augmentation anticipée de l'activité desréseaux intracorticaux inhibiteurs de M1 diminue l'excitabilité corticale, menant à une diminutionde l’excitabilité corticospinale, réduisant ainsi l’amplitude du mouvement évoqué par TMS.Dans les études suivantes (2, 3 et 4), nous nous sommes intéressés aux mécanismescorticaux et corticospinaux impliqués dans la préparation et la réaction rapide à uneperturbation périphérique du mouvement. Nous avons demandé au sujet soit de résister soitde se laisser-faire par une extension passive du poignet, et avons étudié les mécanismesimpliqués dans la modulation de la composante à longue latence du réflexe d'étirement (LLSR,qui débute environ 50 ms après la perturbation), en fonction de l'intention. Concernantl’excitabilité corticospinale, les résultats montrent que, lors de la préparation à uneperturbation périphérique, les phénomènes d'intégration sensori-motrice engendrés par lesafférences sensorielles dues à la perturbation sont pris en compte dans le réglage anticipé del'excitabilité corticospinale, afin que la réaction, déclenchée par les afférences sensorielles, soitadaptée à l'intention du sujet (étude 2). Au niveau cortical, une modification de l'activité desréseaux intracorticaux de M1 en fonction de l'intention précède la modulation de l'activitécorticale du cortex sensorimoteur primaire, liée à la genèse du LLSR, suggérant que desprocessus anticipateurs influencent l’activité du cortex sensorimoteur primaire afin que saréponse précoce à la perturbation soit adaptée à l'intention du sujet (étude 3). Enfin, dansl’étude 4, nous avons mis en évidence le rôle d'une aire motrice non primaire, la SMA proper,dans la modulation du réflexe d'étirement en fonction de l'intention.Ainsi, lorsque nous anticipons une perturbation motrice, des processus préparatoiresspécifiques (dépendants de notre intention), et différents de ceux impliqués avant la réalisationd’un mouvement sans variation des forces externes, sont mis en jeu dans la SMA proper et lecortex sensorimoteur primaire de manière à ce que la réaction rapide, déclenchée au niveau ducortex sensorimoteur par les afférences sensorielles induites par la perturbation, soit adaptée àl’intention du sujet
In everyday life, our actions can be perturbed by rapid variations of environmentalexternal forces. In order to achieve our goals, we have to react “well and fast” to thesemovement perturbations. This reaction implies both cognitive and sensorimotor processes. Wewere interested in the cortical mechanisms (mainly involving the primary motor cortex, M1)underlying the interaction between cognitive and sensorimotor functions that allows theadaptation of the reaction to the perturbation according to the intention. We tried to relate themechanisms implicated during the preparation with those implicated during the realization ofthe reaction. With combined EEG-TMS (with EMG recording), we used a stimulation-recordingapproach, allowing simultaneous observation of cortical and corticospinal mechanisms, bothbefore and after the stimulation. This approach helps to obtain to a better understanding of therelationship between cerebral activity and behavior.In the first experiment, we used a central motor perturbation, i.e. subjects were asked toresist or to assist a movement evoked directly at the cortical level using TMS. We showed thatcognitive processes can directly influence cortical and corticospinal excitability before anyinvolvement of the sensorimotor processes related to the movement execution. When subjectsprepared to resist the TMS-evoked movement, the anticipatory increased activity of theintracortical inhibitory networks of M1 decreased the cortical excitability, leading to adecreased corticospinal excitability and thus to a reduced TMS-evoked movement.In the following experiments (2, 3 and 4), we were interested in cortical andcorticospinal mechanisms engaged during the preparation and the reaction to a peripheralmovement perturbation. We asked subjects either to resist or to not-react (to “let-go”) to apassive wrist extension, and we studied the mechanisms underlying the modulation of the longlatency stretch reflex (LLSR, starting about 50 ms after the perturbation) according to theintention. Concerning the corticospinal excitability, the results showed that, during thepreparation of a reaction to a peripheral perturbation, the anticipatory tuning of thecorticospinal excitability takes into account sensorimotor integrative phenomenons induced bythe afferent input due to the perturbation in such a way that the reaction, triggered by theafferent inputs, is adapted to the subject’s intention (experiment 2). At the cortical level, achange of M1 intracortical network activity (before the perturbation) precedes the modulationof the primary sensorimotor cortex activity that is linked to the LLSR generation (after theperturbation). This strongly suggests that anticipatory processes preset the primarysensorimotor cortex in order to adapt its early response to the perturbation according to thesubject’s intention (experiment 3). Finally, temporary inactivation of SMA proper (induced byTMS) showed that this non-primary motor area is also implicated in the modulation of thestretch reflex according to the intention (experiment 4).In conclusion, when we expect a motor perturbation, intention-specific preparatoryprocesses are engaged in SMA proper and the primary sensorimotor cortex that are differentfrom those involved in the realization of a movement without external force variations. Thesepreparatory processes allow the early motor reaction, generated by the primary sensorimotorcortex (triggered by the afferent input induced by the perturbation) to be adapted to thesubject’s intention

Neuroscience faces the challenging task of developing and implementing objective measures of consciousness that can be applied to patients who are unable to interact with their external environment. The standard clinical assessment of these patients relies heavily on the subjective distinction between voluntary and involuntary or reflexive movements and electrophysiological and neuroimaging protocols have been recently developed to improve diagnosis and probe for signs of awareness. However, because the ability to unambiguously infer the capacity for consciousness through these novel techniques is determined ultimately not by consciousness itself but the awareness of a specific stimulus, their use to diagnose consciousness at the single-patient level is challenged by difficulties related to the application and interpretation of results. This thesis addresses the possibility for investigating the brain’s capacity for consciousness, instead of the neural correlates of particular conscious perceptions, following a path that has not yet been explored. General considerations about what constitutes the content of consciousness led us to hypothesize that consciousness depends on the brain’s capacity to sustain complex patterns of causal interactions between different areas of the thalamocortical system. To investigate this hypothesis, we employed the combination of navigated transcranial magnetic stimulation (TMS) and high-density electroencephalography (hd-EEG) and developed a feasible measure of brain complexity, the Perturbational Complexity Index (PCI), that was calculated in healthy subjects during alert wakefulness, sleep and anesthesia; and at the bedside of brain-injured, non-communicating patients, who gradually recovered from coma. PCI is a measure of the spatiotemporal complexity of the cortical activity evoked by TMS and is high only if many regions of the cerebral cortex react to the initial perturbation quickly and in different ways. Remarkably, in a total of 116 TMS sessions collected from 19 healthy subjects and 17 brain-injured patients, we invariably found high PCI values in conditions in which consciousness was clearly present and low PCI values in conditions in which consciousness was unambiguously reduced. This difference was able to reliably discriminate between conscious and unconscious healthy subjects, producing disjoint distributions that were independent of the stimulation parameters, the strength and the extent of the cortical activation. Moreover, PCI was able to detect progressive changes in consciousness, such as those that occur while a subject is falling asleep, and to discriminate between ambiguous consciousness levels (minimally conscious state) in patients suffering from disorders of consciousness from both lower (vegetative state, sleep/anesthesia) and higher (locked-in syndrome, healthy wakefulness) levels of consciousness. The spatiotemporal complexity of the cortical activity evoked by TMS is a single number that can be calculated at the bedside with little a priori information. Because this measure aims at the brain’s capacity for consciousness, instead of behavioral or neural correlations of conscious perception, this technique does not depend on the willingness or ability of the patient to engage in assessment protocols and can be employed bypassing sensory pathways and subcortical structures to directly probe the thalamocortical system. Our results support PCI as an appropriate tool to approximate an objective measure of the neural correlate of consciousness with the potential to assist the diagnosis and prognosis in brain-injured patients and with unique theoretical implications to a science of consciousness.

Recent experiments showed that sleep SWA can be regulated locally in the cerebral cortex, pointing to a link between SWA regulation and synaptic plasticity, and, more specifically, that local SWA may increase after manipulations that favor local synaptic potentiation and decrease after those that promote local synaptic depression. To further investigate the connection between cortical plasticity and sleep SWA, in the first study presented on this dissertation we employed hd-EEG recordings together with a paired associative stimulation TMS protocol. As expected, such a protocol lead to a sustained increase (LTP-like) or decrease (LTD-like) of cortical excitability as measured by both motor evoked potentials and TMS-evoked cortical responses over sensorimotor cortex. During subsequent sleep, SWA increased locally in subjects whose TMS-evoked cortical responses had increased after PAS, and decreased in subjects whose cortical responses had decreased. Changes in TMS-evoked cortical EEG response and change in sleep SWA were localized to similar cortical regions and were positively correlated. In the second study, a whole night hd-EEG recording was adopted in order to assess the effects of an implicit visuomotor learning task -performed 12 hours before sleep time- over sleep SWA. The predicted increase in SWA was found over a cluster of five electrodes projecting over the right parietal cortex -a region whose circuits are specifically involved during task execution- thus confirming a close relationship between learning processes, synaptic plasticity and local sleep regulation. Together, these results suggest that changes in cortical excitability lead to corresponding changes in local sleep regulation, as reflected by SWA, thus providing evidence for a tight relationship between cortical plasticity and sleep intensity and suggesting a role for sleep in regulating cortical connection strength at synapses as proposed by the Synaptic Homeostasis Hypothesis (Tononi and Cirelli, 2003).

The present thesis comprises two main parts: one theoretical and one experimental. The first part, composed of two chapters, is an in-depth introduction to transcranial magnetic stimulation (TMS) and its simultaneous use with neuroimaging techniques (coregistration). The second part is composed of some of the studies I conducted during my PhD. I chose to include three studies representing the different aspects of my research in the last three years, mainly regarding the study and the application of TMS-EEG coregistration in research (study 1), clinics (study 2) and technical methodology (study 3). The first study (study 1), conducted at the Department of General Psychology of Padova, was aimed to investigate the neuromodulatory effects of an rTMS protocol on healthy volunteers. The second study (study 2) was conducted at the Institute of Neurology of University College London in the context of the international “TrackOnHD” longitudinal project aimed to investigate Huntington disease (HD) in a multimodal approach. The target of this study was to investigate possible TMS-EEG markers of inhibition deficits in Huntington patients. The third study (study 3), conducted in collaboration with the Department of Information Engineering of Padova, was aimed to develop an algorithm of correction to remove an artefact induced by TMS during EEG recordings. CHAPTER I – TRANSCRANIAL MAGNETIC STIMULATION In the last twenty years the development of new techniques able to investigate the brain function in vivo during cognitive and motor tasks lead to impressive advances in understanding the human brain. Transcranial magnetic stimulation (TMS) is a tool whose popularity has grown progressively thanks to its ability to stimulate the brain in a focal and non-invasive way (Barker et al., 1985), permitting to establish a causal link in the brain-cognition/motor-behaviour relationship (Pascual-Leone et al., 2000). In the first chapter of this thesis the possible applications of TMS in the field of cognition, physiology and rehabilitation are discussed. Specifically, the first part focuses on the operating mechanisms of TMS and on the different stimulation parameters that define the effects of the stimulation. In the second part of the first chapter, the three main TMS protocols are discussed: single-pulse TMS, which is used in the temporal and spatial characterization of cognitive processes, in the study of motor cortex reactivity, and in the investigation of the cortico-spinal tract functioning; paired-pulse TMS, that investigates the connectivity and the interaction of cerebral networks at rest or during a task performance; and repetitive TMS (rTMS), that explores the cerebral plasticity processes both in relation to cognitive processing and for rehabilitation treatments. CHAPTER II – THE SIMULTANEOUS USE OF TRANSCRANIAL MAGNETIC STIMULATION WITH EEG AND OTHER NEUROIMAGING TECHNIQUES Despite the widespread use of TMS in current research, its mechanism of action is still poorly understood (Miniussi et al., 2010). This lack in comprehension results from missing a firsthand “visible” marker of cortical response and a need for secondary measures of primary motor and visual cortex stimulation. In the last twenty years, thanks to the progressive improvements in neuroimaging technology, the first attempts to simultaneously use TMS with other neuroimaging techniques have been made possible (e.g. TMS-EEG, Ilmoniemi et al., 1997; TMS-PET, Paus et al., 1997). On one hand, the possibility to actively stimulate the brain with TMS allows to establish “causal” inferences in neuroimaging studies, in which, traditionally, only “correlational” inferences were possible. On the other hand, neuroimaging techniques potentially provide an important contribution through the spatial and temporal information of the neural activation evoked by TMS. In the second chapter of this thesis, the strong and the weak points of different TMS-neuroimaging coregistration approaches are depicted. Specifically, the middle part of the chapter focuses on the main topic of this thesis, i.e. the TMS-EEG coregistration. TMS-EEG, among the different approaches, is the most successful and widespread, thanks to its promising value in the investigation of brain dynamics. Indeed, EEG is able to record the post-synaptic potentials following the neuronal depolarization evoked by TMS at a high temporal resolution (Ilmoniemi et al., 1997). The analysis of the TMS-evoked EEG activity in terms of time, space, frequency and power, potentially provides important and accurate information in the local activation induced by the stimulation (cerebral reactivity), in the spread of such activation (cerebral connectivity), and in the long-lasting neuromodulatory effects following rTMS protocols (cerebral plasticity). On the other hand, the TMS-EEG coregistration, presents several technical difficulties mainly due to the different artefacts that electromagnetic stimulation induces in the EEG signal. These aspects are discussed thoroughly in the second chapter. Finally, the last part of the second chapter is dedicated to the other TMS coregistration approaches with magnetic resonance imaging (MRI), functional magnetic resonance imaging (fMRI), positron emission tomography (PET), single-photon emission computed tomography (SPECT) and near-infrared spectroscopy (NIRS). CHAPTER III – STUDY 1: NEUROMODULATORY EFFECTS OF LOW-FREQUENCY RTMS: INSIGHTS FROM TMS-EEG The neuromodulatory effects of rTMS have been mostly investigated by means of peripheral motor-evoked potentials (MEPs). However, MEPs are an indirect measure of cortical excitability, also being affected by spinal excitability. The development of new TMS-compatible EEG systems allowed the direct investigation of the stimulation effects through the cortical responses evoked by TMS (TEPs). In this study, we investigated the effects of a repetitive TMS (rTMS) protocol delivered at low frequency (1 Hz), which is known to produce an inhibitory effect on cortical excitability (Chen et al., 1997). The protocol was applied over the primary motor cortex of 15 healthy volunteers and, as a control, over the primary visual cortex of 15 different healthy volunteers to examine the spatial specificity of the stimulation. The effects of the stimulation were analyzed in both groups through the single-pulse stimulation of the primary motor cortex, before and immediately after the rTMS protocol. Different measures were tested: MEPs, TEPs, local mean field power and scalp maps of the activity distribution. Results on MEPs amplitude showed a significant reduction following the rTMS over the primary motor cortex. Results on TEPs, showed a well-known TEPs pattern evoked by single-pulse stimulation of the motor cortex: P30, N45, P60 and N100. Following the motor cortex rTMS, we observed a significant increase of P60 and N100 amplitude, whose origin has been linked to the GABAb-mediated inhibitory post-synaptic potentials (Ferreri et al., 2011; Premoli et al, 2014). Results on LMFP, showed an increase of general activity induced by the single-pulse stimulation of the motor cortex, starting from 90 ms after the TMS pulse. This latency actually corresponds to the peak of GABAb inhibition. No significant effects were detected after rTMS of the primary visual cortex. The results of this study are relevant in three main aspects: (1) we confirmed the inhibitory effect of 1-hz rTMS, also providing a central correlate of such effect (TEPs); (2) we defined the spatial specificity and the origin of the inhibitory effect of 1-Hz rTMS; (3) we confirmed the possible role of the TMS-evoked N100 as a cortical inhibitory marker. The present findings could be of relevance both for therapeutic purposes, especially for pathologies characterized by inhibitory deficits (e.g. Parkinson’s disease; Huntington’s disease); and for basic research, especially in studies aimed to correlate a behavioral performance to the amount of cerebral excitability. CHAPTER IV – STUDY 2: TMS-EEG MARKERS OF INHIBITORY DEFICIT IN HUNTINGTON’S DISEASE Recent studies have shown the potential value of combining TMS and EEG for clinical and diagnostic purposes. Several TMS-EEG measures in terms of evoked potentials (i.e. TEPs), brain sources analysis, oscillatory activity and global power has been used in the assessment of brain dynamics deficits in several pathologies, such as: schizophrenia (Ferrarelli et al., 2008); psychotic disorders (Hoppenbrouwers et al., 2008); depression (Kähkönen et al., 2005); awareness disorders (Massimini et al., 2005); epilepsy (Rotenborg et al., 2008) and autism (Sokhadze et al., 2012). For instance, the potential contribution of TEPs in the investigation of the cerebral facilitatory/inhibitory balance has been demonstrated, given their origin from different GABAergic neuronal populations (Ferreri et al., 2011; Premoli et al., 2014). In particular, the TMS-evoked N100 has been related to the amount of GABAergic inhibition, as shown by pharmacological (e.g. Kähkönen et al., 2003) and behavioral research (e.g. Bender et al., 2005; Bonnard et al., 2009) as well as studies in patients (e.g. Helfrich et al., 2013). As a part of the multi-site international “TrackOnHD” project, we used TMS-EEG to investigate the electrophysiological markers of motor cortex stimulation in Huntington patients. In Huntington’s disease (HD) the progressive degeneration of GABAergic neurons in the striatum lead to a strong reduction of inhibition, resulting in an excessive increase in glutamatergic excitability (i.e. excitoxicity). Our study compared a group of 12 HD patients with a group of 12 healthy volunteers over several different TMS-EEG, EMG, fMRI and clinical measures (in the chapter only the TMS-EEG results are reported). We found a specific and significant decrease of the N100 as assessed by the time point-by-time point permutation analysis of TEPs and from the analysis of the global activity from 90 to 104 ms after the TMS pulse. Scalp maps of the activity distribution showed a bilateral decrease of negativity, such effect was stronger over the site of stimulation. Event-related spectral perturbation and inter-trial coherence analysis showed a significant difference in the oscillatory activity of the two groups within the GABAb-ergic time window (i.e. 60-110 ms after the TMS pulse). We speculated that the observed results might be produced by the deficit in GABAergic inhibition as a consequence of the striatum neuronal degeneration in HD patients. Although preliminary, these results provided potentially useful TMS-EEG markers for inhibitory deficits in HD patients. Further analyses are needed to correlate the present findings with the other measures collected. CHAPTER V – STUDY 3: TMS-EEG ARTIFACTS: A NEW ADAPTIVE ALGORITHM FOR SIGNAL DETRENDING During EEG recording the discharge of the TMS coil may generate an artefact that can last for tens of milliseconds, known as “decay artefact” (Rogasch et al., 2014). This can represent a problem for the analysis of the TMS-evoked potentials (TEPs). So far, two main strategies of correction have been proposed involving the use of a linear detrend or independent component analysis (ICA). However, none of these solutions may be considered optimal: firstly, because in most of the cases the decay artefact shows a non-linear trend; secondly, because the ICA correction (1) might be influenced by individual researcher’s choices and (2) might cause the removal of physiological responses. Our aim is to verify the feasibility of a new adaptive detrend able to discriminate the different trends of the decay (linear or non-linear). Forty healthy volunteers were stimulated with 55 TMS pulses over the left M1. The TMS-EEG responses were compared among five conditions: RAW (no correction of the decay artefact was applied); INFOMAX29 (the decay components were extracted and removed by the INFOMAX ICA algorithm, using 31 electrodes); FASTICA (the decay components were extracted and removed by the fastICA ICA algorithm, using 31 electrodes); INFOMAX15 (the decay components were extracted and removed by the INFOMAX ICA algorithm, using 15 electrodes) and ALG (the decay artefact was corrected through the use of an adaptive algorithm). To assess whether the artefact correction significantly affected the physiological responses to TMS as well, we examined the differences in the -100 + 400 ms time window around the TMS pulse by means of a non-parametric, cluster-based, permutation statistical test. Then we compared the peak-to-peak TEPs amplitude within the detected time windows. The grand-averaged EEG response revealed five main peaks: P30, N45, P60, N100 and P180. Significant differences (i.e. Monte Carlo p-values < 0.05) were detected in a cluster nearby the TMS coil, and specifically over FC1, CP1, C3 and FC2. Repeated-measures ANOVA revealed a significant corruption of the peak-to-peak amplitude after INFOMAX29 (3 TEPs out of 8), FASTICA (4 TEPs out of 12), INFOMAX15 (5 TEPs out of 15) and ALG correction (2 TEPs out of 15), compared to the original signal. Furthermore, abnormal LMFP and TEPs scalp distribution were detected following the INFOMAX29 and FASTICA correction. When our algorithm was used, however, the TEPs amplitude, morphology and distribution was in line with the literature and not significantly different from the original signal. Also the decay artefact was correctly removed. The main contribution of this study is the proposal of a new adaptive algorithm to correct the decay artefact induced by TMS in the EEG signal. Our results demonstrated that the proposed adaptive detrend is a reliable solution for the correction of this artefact, especially considering that, contrary to ICA, (1) it is not dependent from the number of recording channels; (2) it does not affect the physiological responses and (3) it is completely independent from the experimenter’s choices.
La presente tesi si compone di due parti principali: una teorica e una sperimentale. La prima parte, suddivisa in due capitoli, è un approfondimento teorico sullo strumento stimolazione magnetica transcranica (TMS) e sul suo utilizzo simultaneo (ossia, in coregistrazione) con le tecniche di neuroimaging. La seconda parte comprende alcuni degli studi condotti durante il mio dottorato. Nello specifico, si tratta di tre studi che coprono i diversi aspetti applicativi delle ricerche che ho condotto in questi tre anni, ossia lo studio e l’utilizzo della coregistrazione TMS-EEG in ricerca (studio 1), in ambito clinico (studio 2) e per aspetti tecnico-metodologici (studio 3). Il primo studio (studio 1), condotto nel Dipartimento di Psicologia Generale di Padova, era volto all’analisi degli effetti neuromodulatori di un protocollo rTMS su volontari sani. Il secondo studio (study 2) è stato condotto all’Istituto di Neurologia dello University College London (Londra, Regno Unito) all’interno del progetto internazionale “TrackOnHD”, uno studio longitudinale avente come obiettivo l'indagine approfondita della Malattia di Huntington (HD) attraverso un approccio multimodale. L’obiettivo di questo studio era la ricerca di potenziali marker TMS-EEG che riflettessero il deficit di inibizione cerebrale che caratterizza questa patologia. Il terzo studio (study 3), svolto in collaborazione col Dipartimento di Ingegneria dell’Informazione di Padova, aveva l’obiettivo di sviluppare un algoritmo di correzione in grado di rimuovere un artefatto indotto dalla TMS durante la registrazione EEG. CAPITOLO I - LA STIMOLAZIONE MAGNETICA TRANSCRANICA (TMS) Negli ultimi anni lo sviluppo di nuove tecniche in grado di analizzare l’attivazione cerebrale durante processi cognitivi e motori, ha portato ad un avanzamento progressivo delle conoscenze sul cervello umano. La stimolazione magnetica transcranica (TMS) è stata uno degli strumenti la cui popolarità è cresciuta in questi ultimi anni, grazie alla possibilità di stimolare, in modo focale e non invasivo, il cervello in vivo (Barker et al., 1985). Tale capacità ha consentito, per la prima volta, la straordinaria possibilità di inferire delle relazioni causali tra cervello, processi cognitivi e motori, e comportamento (Pascual-Leone et al., 2000). Nel primo capitolo della presente tesi vengono passate in rassegna tutte le possibili applicazioni della TMS in campo cognitivo, fisiologico e riabilitativo. Nello specifico, la prima parte è dedicata ai meccanismi di funzionamento della TMS e ai parametri di stimolazione che ne definiscono i diversi effetti sul cervello. Nella seconda parte vengono invece passati in rassegna i tre principali protocolli di stimolazione: la TMS a singolo impulso, utilizzata per la caratterizzazione spaziale e temporale dei processi cognitivi, per analizzare la reattività della corteccia motoria primaria, e per verificare l’integrità del tratto cortico-spinale; la TMS a doppio impulso, per studiare la connettività e l’interazione di network cerebrali a riposo e durante lo svolgimento di un task; e la TMS ripetitiva (rTMS), utilizzata per analizzare i fenomeni di plasticità cerebrale sia durante processi cognitivi, sia in relazione a trattamenti riabilitativi. CAPITOLO II - L’UTILIZZO SIMULTANEO DELLA TMS CON L’EEG ED ALTRE TECNICHE DI NEUROIMAGING Nonostante la grande popolarità che la TMS ha conosciuto negli ultimi anni, molti aspetti del suo meccanismo d’azione sono ancora poco chiari (Miniussi et al., 2010). Tale ambiguità è dovuta al fatto che, fatta eccezione per la corteccia motoria e visiva primaria, la stimolazione TMS non fornisce dei marker “visibili” di eccitabilità corticale. Negli ultimi anni, grazie al miglioramento tecnologico degli strumenti di indagine neuroscientifica, si è iniziato a utilizzare simultaneamente (in coregistrazione) la TMS con diverse tecniche di neuroimaging. Ciò ha consentito di trarre delle inferenze di tipo “causale” e non più solo “correlazionale” (come nei tradizionali studi di neuroimaging) grazie alle informazioni spaziali e temporali sull’effetto della TMS che le tecniche di neuroimaging offrono. Nel secondo capitolo della presente tesi, vengono trattati dettagliatamente le potenzialità e i limiti delle diverse coregistrazioni TMS-neuroimaging. In particolare, nella parte centrale del capitolo è dato ampio spazio all’argomento centrale di questa tesi, ossia la coregistrazione TMS-EEG. L’approccio TMS-EEG, tra i vari metodi di coregistrazione, è stato quelli che negli ultimi anni ha riscontrato maggiore successo e diffusione, dovuto all’enorme potenzialità che questo metodo garantisce nello studio delle dinamiche cerebrali. L’EEG, infatti, è in grado di registrare, ad altissima risoluzione temporale, i potenziali post-sinaptici indotti dalla depolarizzazione neuronale evocata dalla TMS (Ilmoniemi et al., 1997). L’analisi dell’attività EEG indotta dalla TMS - in termini di tempo, spazio, frequenza e potenza - è in grado di fornire delle preziose informazioni sia sull’attivazione locale indotta dalla stimolazione (reattività cerebrale), sia su quella distale (connettività cerebrale), sia sulle modificazioni a seguito di protocolli di stimolazione ripetitiva (plasticità cerebrale). D’altra parte, la coregistrazione TMS-EEG presenta numerose difficoltà di tipo tecnico, dovuto ai numerosi artefatti che la stimolazione elettromagnetica induce sul segnale EEG (così come sui segnali delle altre tecniche di neuroimaging), questi aspetti sono trattati in maniera dettagliata all’interno del capitolo. Infine, l’ultima parte del capitolo è dedicata agli altri metodi di coregistrazione TMS con risonanza magnetica (MRI), risonanza magnetica funzionale (fMRI), tomografia a emissione di positroni (PET), tomografia a emissione di fotone singolo (SPECT) e spettroscopia del vicino infrarosso (NIRS). CAPITOLO III – STUDIO 1: EFFETTI NEUROMODULATORI DELLA RTMS A BASSA FREQUENZA: EVIDENZE DALL’APPROCCIO TMS-EEG Tradizionalmente gli effetti neuromodulatori della rTMS sono stati studiati attraverso l’analisi dei potenziali motori evocati (MEP). Tuttavia, come noto, i MEP sono una misura indiretta dell’eccitabilità corticale avendo una forte componente anche spinale. Con lo sviluppo di nuovi sistemi EEG compatibili con la TMS, è stato possibile analizzare gli effetti della stimolazione in modo più diretto, tramite l’analisi dei potenziali corticali evocati dalla TMS (TEPs). In questo studio abbiamo analizzato l’effetto di un protocollo di TMS ripetitiva (rTMS) a bassa frequenza (1 Hz) molto noto, soprattutto in ambito riabilitativo, per sortire un effetto di inibizione dell’eccitabilità corticale. Il protocollo è stato applicato sulla corteccia motoria primaria di quindici volontari sani e sulla corteccia visiva primaria di altri quindici volontari sani, assunti come gruppo di controllo per analizzare la specificità spaziale della stimolazione. Gli effetti della stimolazione ripetitiva sono stati testati su diverse misure elettrofisiologiche evocate da una stimolazione a singolo impulso della corteccia motoria, prima e subito dopo il protocollo rTMS, ossia: MEP, TEPs, local mean field power (LMFP) e distribuzione dell’attività sullo scalpo. I risultati sui MEP hanno mostrato una diminuzione significativa dell’ampiezza a seguito del protocollo rTMS sulla corteccia motoria. I risultati sui TEP hanno mostrato un pattern noto composto di quattro principali picchi: P30, N45, P60 e N100. A seguito del protocollo rTMS sulla corteccia motoria si è osservato un incremento significato dell’ampiezza dei TEP P60 e N100, la cui origine è legata all’attività dei potenziali post-sinaptici inibitori GABAb (Ferreri et al., 2011; Premoli et al., 2014). I risultati sul LMFP hanno mostrato un incremento di attività generale indotta dalla TMS sulla corteccia motoria a partire da circa 90 ms dalla stimolazione, ossia la latenza del picco massimo di inibizione GABAb. A seguito del protocollo di stimolazione di controllo, applicato sulla corteccia visiva, non si è riscontrato nessun cambiamento significativo. I risultati di questo studio hanno una rilevanza su tre aspetti: (1) si è confermato l’effetto inibitorio del protocollo rTMS a 1-Hz, offrendo anche un correlato centrale di inibizione (TEPs) oltre che periferico (MEPs); (2) sono state definite la spazialità e l’origine dell’inibizione indotta dalla rTMS a bassa frequenza; (3) la N100 evocata dalla TMS si conferma essere un marker affidabile del grado di inibizione corticale. I risultati di questo studio potrebbero avere una rilevanza sia in campo terapeutico e riabilitativo, specie per i disturbi alla cui base si suppone vi sia un deficit di inibizione corticale (ad es. malattia di Parkinson, malattia di Huntington); sia in campo di ricerca, specie in studi in cui si vogliano correlare performance a task cognitivi o motori con il grado di eccitazione/inibizione corticale. CAPITOLO IV – STUDIO 2: DEFICIT DI INIBIZIONE NELLA MALATTIA DI HUNTIGTON: EVIDENZE DALLA COREGISTRAZIONE TMS-EEG Evidenze recenti hanno mostrato le potenzialità dell’utilizzo della coregistrazione TMS-EEG in ambito clinico e diagnostico. Diverse misure TMS-EEG in termini di potenziali evocati (TEPs), analisi di sorgenti, attività oscillatoria e potenza dell’attività globale, sono state utilizzate per lo studio di dinamiche cerebrali deficitarie in diverse patologie, come: schizofrenia (Ferrarelli et al., 2008); disordini psicotici (Hoppenbrouwers et al., 2008); depressione (Kähkönen et al., 2005); disturbi di coscienza (Massimini et al., 2005); epilessia (Rotenborg et al., 2008) e autismo (Sokhadze et al., 2012). Ad esempio, diverse evidenze hanno mostrato il potenziale contributo dei TEPs nello studio degli equilibri eccitatori/inibitori corticali, data la loro origine GABAergica (Ferreri et al., 2011; Premoli et al., 2014). In particolare, la N100 TMS-evocata sembra essere strettamente correlata al grado di inibizione GABAergica, come mostrato da evidenze a carattere farmacologico (ad es. Kähkönen et al., 2003; Premoli et al., 2014); studi comportamentali (ad es. Bender et al., 2005; Bonnard et al., 2009) e studi in pazienti (ad es. Helfrich et al., 2013). Nel presente studio, facente parte di un ampio progetto internazionale multicentrico (“TrackOnHD”), abbiamo utilizzato la coregistrazione TMS-EEG per analizzare dei possibili marker elettrofisiologici della malattia di Huntington, tramite stimolazione della corteccia motoria primaria. La malattia di Huntington (HD) è caratterizzata da una progressiva degenerazione dei neuroni striatali di natura GABAergica. Tale degenerazione porta a un eccessivo incremento del tono eccitatorio mediato dal glutammato, un fenomeno noto come eccitossicità. Nel presente studio sono stati analizzati dodici pazienti HD e dodici volontari sani su varie misure TMS-EEG, EMG, fMRI e cliniche (nel capitolo sono riportati solo i risultati relativi alle misure TMS-EEG). I risultati hanno mostrato una riduzione significativa e specifica della N100, come rilevato dall’analisi dei TEP per permutazioni punto-per-punto e dall’analisi dell’attività media globale da 94 a 104 ms dopo l’impulso TMS. Le mappe dello scalpo della distribuzione dell’attività hanno mostrato una riduzione della negatività su entrambi gli emisferi, con un effetto maggiore sul sito di stimolazione. Le analisi di perturbazione dello spettro evento-relata e della coerenza inter-trial hanno mostrato una differenza significativa nell’attività oscillatoria dei due gruppi all’interno della finestra di interesse GABAb-ergico (60-110 ms dopo l’impulso TMS). I risultati osservati potrebbero essere prodotti dal deficit di inibizione GABAergica nei pazienti HD conseguente alla degenerazione neuronale nello striato. Anche se preliminari, i risultati dello studio hanno rilevato dei marker TMS-EEG potenzialmente d’interesse per la valutazione dei deficit inibitori in pazienti HD. Ulteriori analisi sono necessarie per correlare i risultati ottenuti con le altre misure raccolte all’interno del progetto. CAPITOLO V – STUDIO 3: ARTEFATTI TMS-EEG: UN NUOVO ALGORITMO ADATTATIVO PER IL DETREND DEL SEGNALE Durante un EEG, la stimolazione TMS può generare un artefatto a lunga latenza, noto come artefatto “decay”. Tale artefatto rappresenta un problema per l’analisi dei potenziali evocati dalla TMS (TEP). In letteratura, per risolvere il problema, sono comunemente utilizzate due principali strategie: l’utilizzo di un detrend lineare e l’utilizzo dell’independent component analysis (ICA). Tuttavia, nessuna di queste soluzioni può essere considerata ottimale. Per quanto riguarda l’utilizzo di un detrend lineare, dal momento che nella maggior parte dei casi l’artefatto decay non segue un andamento lineare, questo tipo di correzione risulta inefficiente. Per quanto invece riguarda l’ICA, anche questa procedura presenta dei limiti intrinseci: (1) può essere eccessivamente influenzato dalle scelte dello sperimentatore e (2) può causare la rimozione di componenti fisiologiche, oltre che artefattuali. Il nostro obiettivo è di verificare l’efficienza di un nuovo detrend adattivo, sviluppato su MATLAB, in collaborazione col dipartimento di Ingegneria Informatica di Padova, capace di discriminare i diversi trend dell’artefatto decay (ossia lineare e non-lineare). Quaranta volontari sani sono stati stimolati con 55 impulsi TMS singoli sulla corteccia motoria primaria di sinistra. Le risposte EEG indotte dalla TMS sono state analizzate in cinque condizioni: RAW (in cui non veniva applicata nessuna correzione dell’artefatto decay); INFOMAX29 (in cui l’artefatto decay veniva corretto con un algoritmo ICA-INFOMAX, considerando tutti i 29 canali); FASTICA (in cui l’artefatto decay veniva corretto con un algoritmo fastICA, considerando tutti i 29 canali); INFOMAX15 (in cui l’artefatto decay veniva corretto con un algoritmo ICA-INFOMAX, considerando solo 15 canali) e ALG (in cui l’artefatto decay veniva corretto tramite il nostro algoritmo adattivo). Per verificare se la correzione dell’artefatto avesse influenzato anche i TEP, sono state analizzare le differenze in una finestra temporale da -100 a +400 ms dall’impulso TMS attraverso l’utilizzo di un test per permutazioni, non-parametrico e corretto per cluster. Successivamente, sono state comparate le ampiezze e le latenze picco-picco dei TEP all’interno delle finestre temporali negli elettrodi risultati significativi. La risposta grand-average ha rilevato cinque picchi principali: P30, N45, P60, N100 e P180. Sono state rilevate delle differenze significative (i.e. Monte Carlo p < 0.05) in un cluster di elettrodi vicino alla stimolazione, comprendente i canali FC1, CP1, C3 e FC2. Le analisi sull’ampiezza picco-picco hanno rilevato una significativa modulazione dell’ampiezza dopo la correzione INFOMAX29 (in 3 TEP su 8), FASTICA (in 4 TEP su 12), INFOMAX15 (in 5 TEP su 15) e ALG (in 2 TEP su 15), rispetto al segnale RAW originale. I risultati LMFP e delle mappe di distribuzione sullo scalpo hanno rilevato diverse anomalie a seguito della correzione INFOMAX29 e FASTICA. I risultati hanno mostrato che la correzione ICA modifica in modo significativo l’ampiezza, la morfologia e la distribuzione di una parte dei TEP analizzati e nello stesso tempo non garantisce una completa rimozione dell’artefatto decay. Al contrario, a seguito della correzione col nostro algoritmo (condizione ALG), l’ampiezza, la morfologia e la distribuzione dei TEP rimanevano fedeli a quella originale, con una rimozione pressoché completa dell’artefatto decay. Il principale contributo di questo studio è stato la proposta di un nuovo algoritmo di correzione per un artefatto a lunga latenza che la TMS induce sul segnale EEG (artefatto decay) rendendo difficoltosa l’analisi. I risultati hanno dimostrato che questo metodo è più efficiente delle strategie attualmente in utilizzo in letteratura, non avendo i limiti intrinseci presentati dall’algoritmo ICA.

Peripheral nerve stimulation (PNS) is commonly used to promote use-dependent cortical plasticity for rehabilitation of motor function in spinal cord injury. Pairing transcranial magnetic stimulation (TMS) with PNS has been shown to increase motor evoked potentials most when the two stimuli are timed to arrive in the cortex simultaneously. This suggests that a mechanism of timing-dependent plasticity (TDP) may be a more effective method of promoting motor rehabilitation. The following thesis is the result of applying a brain-computer interface to apply PNS in closed-loop simultaneously to movement intention onset as measured by EEG of the sensorimotor cortex to test whether TDP can be induced in incomplete spinal cord injured individuals with upper limb motor impairment. 4 motor incomplete SCI subjects have completed 12 sessions of closed-loop PNS delivered over 4-6 weeks. Benefit was observed for every subject although not consistently across metrics. 3 out of 4 subjects exhibited increased maximum voluntary contraction force (MVCF) between first and last interventions for one or both hands. TMS-measured motor map volume increased for both hemispheres in one subject, and TMS center of gravity shifted in 3 subjects consistent with studies in which motor function improved or was restored. These observations suggest that rehabilitation using similar designs for responsive stimulation could improve motor impairment in SCI.

Autism spectrum disorder (ASD) is a pervasive neurodevelopmental condition that affects almost 1% of the population. One of the main challenges in the current ASD research is to define the early neurocognitive impairments that provide critical foundations for the core deficits in social and communication abilities. In particular, early attentional dysfunctions may play a critical role in the emergence of ASD. In this doctoral thesis I present six studies that give significant new insights into the nature of altered visual attention in individuals with ASD and their possible neural underpinnings. In the first study we show that individuals with ASD are impaired in enlarging (i.e., “zooming-out”) the attentional focus size relative to the control group and this deficit can impact the rapid orienting toward a cued location in the visual field. The second and the third studies show how parents without any history of ASD but with elevated autistic traits can transmit to their infants subtle deficit in visual attention (at the expense of both orienting and zooming mechanism) that may impact children’s future socio-communicative abilities. In the fourth and the fifth studies we employed transcranial magnetic stimulation and dense-array electroencephalography, respectively, with typical adults participants and we show that a network of frontal (mainly FEF and IFG) and parietal (mainly IPS/SPL) brain areas are fundamental in regulating the size of the attentional focus. In the last study, we evaluated the spatial profile of the attentional focus in individuals with ASD and results show that the inhibitory ring outside the focus of attention – fundamental to attenuate processing of irrelevant information – is significantly weakened relative to the control group. Overall, these findings show the importance of attentional impairments in the core manifestations of ASD and in its developmental course. Defining attentional abnormalities and their neural correlates is extremely important (i) to improve the early detection of the disorder and, (ii) to implement timely prevention programs to reduce the incidence of ASD.
Il disturbo dello spettro autistico (DSA) è un disturbo neuroevolutivo pervasivo che colpisce quasi l'1% della popolazione. Una delle principali sfide nell'attuale ricerca sul DSA è definire i deficit neurocognitivi precoci che costituiscono le fondamenta dei disturbi "chiave" nelle abilità sociali e comunicative. In particolare, precoci disfuzioni attentive potrebbero giocare un ruolo decisivo nell'emergere del DSA. Nella presente tesi di dottorato presento sei studi che contribuiscono significativamente alla comprensione delle alterazioni dell'attenzione visiva nei DSA e le loro possibili basi neurali. Nel primo studio, mostriamo che gli individui affetti da DSA sono compromessi nell'abilità di allargare ("zoom-out") la dimensione del fuoco attentivo e che questo problema può avere un impatto negativo nell'orientamento rapido verso una posizione segnalata nel campo visivo. Il secondo e terzo studio mostrano come genitori senza alcuna storia clinica di DSA ma con elevati tratti autistici possano trasmettere ai loro infanti sottili alterazioni nell'attenzione visiva (a carico sia del meccanismo di orientamento che di quello di zoom) che possono avere conseguenze negative sul futuro sviluppo delle abilità socio-comunicative dei loro figli. Nel quarto e quinto studio, abbiamo utilizzato la stimolazione magnetica transcranica e l'elettroencefalografia ad alta densità, rispettivamente, in partecipanti adulti a sviluppo tipico e mostriamo che un network di aree frontali (principalmente FEF e IFG) e parietali (principalmente IPS/SPL) sono fondamentali nella regolazione della dimensione del fuoco attentivo. Nell'ultimo studio, abbiamo valutato il profilo spaziale del fuoco attentivo in individui con DSA e mostriamo come l'anulo inibitorio circostante al fuoco attentivo – fondamentale per attenuare il processamento d'informazioni irrilevanti – è significativamente più debole nel DSA rispetto al gruppo di controllo. Complessivamente, queste evidenze mostrano l'importanza dei deficit attentivi nelle manifestazioni chiave del DSA e nel suo decorso evolutivo. Definire le anomalie dell'attenzione e i corrispondenti correlati neurali è estremamente importante (i) per migliorare la diagnosi precoce del disturbo e (ii) per implementare tempestivi programmi preventivi mirati a ridurre l'incidenza dei DSA.

A few minutes at a busy square in London allow one to appreciate the wide array of actions that humans are capable of expressing— walking, reading a book, tapping touch screen of smartphone, eating, shaking hands and crossing the street. Response inhibition is an essential mechanism of action control and is one of the most studied processes. For example, crossing the street when a fast motorcycle is approaching might necessitate inhibition of stepping forward to avoid being hurt. This ability to quickly suppress a response in a dynamic environment has traditionally been associated with conscious control. Crucially, recent experimental evidence has challenged the view that inhibitory control is restricted to conditions where stimuli are accessible to conscious awareness. Such an unconscious and automatic activation of the motor response system does not necessarily require stimuli to be consciously perceived and is deemed essential to act in a constantly changing environment. This has been interpreted as a basic motor process allowing preparatory mechanisms to automatically suppress an activated movement without the need of conscious cognitive processes. Thus, while there may be differences between automatic and voluntary processes, they might not have entirely distinct neural representations. Indeed, automatic control appears to rely on the corticobasal ganglia network that has been associated with voluntary control. Contemporary research has shown that an up-regulation of neural beta oscillations in the cortico-basal ganglia dynamics can be functionally relevant for inhibition of movement. Consequently, beta oscillations have been proposed as an essential mechanism that allows the motor network to communicate in a dynamic and flexible manner. Present research has demonstrated that it is possible to interact with the neuronal activity by non invasive brain stimulation (NIBS) techniques such as transcranial Direct Current Stimulation (tDCS), transcranial Alternating Current Stimulation (tACS). Specifically, tACS allows delivery of alternating current at different frequencies and it has been used to manipulate ongoing brain oscillations in a controllable way. This concept is still in the very early stages of research, and much needs to be done in order to fully grasp the underlying mechanisms. Building upon these discoveries, the research presented in this thesis aimed to demonstrate a causal role of beta frequency oscillations on unconscious and automatic inhibition adopting tACS over the primary motor cortex and supplementary motor area. Furthermore combining tACS with TMS and EEG allowed me to characterise the underlying basic mechanisms of its action on corticospinal excitability and neuronal dynamics. Overall, this work contributes to our understanding of the human motor system while offering new insights into the combined approach of tACS and EEG in the characterization of a causal role of neuronal oscillatory dynamics on behaviour.
Alcuni minuti in una piazza affollata di Londra permettono di apprezzare l'ampia gamma di azioni che gli esseri umani sono capaci di esprimere— camminare, leggere un libro, toccare lo schermo dello smartphone, mangiare, stringere la mano e attraversare la strada. L'inibizione della risposta è un meccanismo essenziale del controllo motorio dell'azione e rappresenta uno dei processi più studiati. Ad esempio, attraversare la strada quando inavvertitamente si avvicina una motocicletta a grande velocità potrebbe richiedere l'inibizione di mettere i piedi giù per evitare di essere feriti. Questa capacità di sopprimere rapidamente una risposta in un ambiente dinamico è stata tradizionalmente associata al controllo cosciente. In modo cruciale, recenti prove sperimentali hanno sfidato la concezione che il controllo inibitorio è limitato alle condizioni in cui gli stimoli sono accessibili alla consapevolezza cosciente. Tale attivazione inconscia e automatica del sistema motorio non necessariamente richiede che gli stimoli siano consapevolmente percepiti e si ritiene essenziale per agire in un ambiente in costante evoluzione. Questa attivazione è stata interpretata come un processo motorio basale che permette a meccanismi preparatori di sopprimere automaticamente un movimento attivato senza la necessità di processi cognitivi coscienti. Così, sebbene ci siano delle differenze tra i processi automatici e quelli volontari, tali processi potrebbero non avere rappresentazioni neurali completamente distinte. Infatti, il controllo motorio automatico sembra avere come substrato neurale il circuito corticale-ganglio basale che è stato associato al controllo motorio volontario. La ricerca contemporanea ha inoltre dimostrato che l’incremento delle oscillazioni beta nelle dinamiche del sistema corticale-ganglio basale può essere funzionalmente rilevante per l'inibizione del movimento. Di conseguenza, le oscillazioni beta sono state proposte come un meccanismo essenziale che consente al network motorio di comunicare in modo dinamico e flessibile. Nel frattempo, la ricerca attuale ha dimostrato che è possibile interagire con l'attività neuronale mediante tecniche di stimolazione cerebrale non invasiva (NIBS) come la stimolazione transcranica a corrente diretta (tDCS), la stimolazione transcranica a corrente alternata (tACS). In particolare, tACS consente la diffusione di corrente alternata a diverse frequenze ed è stata utilizzata per manipolare le oscillazioni cerebrali in modo controllabile. Comunque, questo concetto è ancora nelle fasi iniziali della ricerca e molto deve essere fatto per comprendere appieno i meccanismi sottostanti. Basandosi su queste scoperte, la ricerca presentata in questa tesi ha lo scopo di dimostrare un ruolo causale delle oscillazioni neurali beta sull’ inibizione inconscia e automatica, adottando la tACS sulla corteccia motoria primaria e l'area motoria supplementare. Inoltre, la combinazione di tACS con TMS e EEG mi ha permesso di caratterizzare i meccanismi di base della sua azione attraverso la misurazione dell’eccitabilità corticospinale e delle dinamiche oscillatorie neuronali. Nel complesso, questo lavoro contribuisce alla nostra comprensione del sistema motorio umano, offrendo al tempo stesso nuove conoscenze sull'approccio combinato di tACS e EEG nella caratterizzazione di un ruolo causale delle dinamiche oscillatorie neuronali sul comportamento.

Il sonno e la veglia vengono comunemente considerati come due stati distinti. L’alternanza tra essi, la cui presenza è stata dimostrata in ogni specie animale studiata fino ad oggi, sembra essere una delle caratteristiche che definisce la nostra vita. Allo stesso tempo, però, le scoperte portate alla luce negli ultimi decenni hanno offuscato i confini tra questi due stati. I meccanismi del sonno hanno sempre affascinato i neurofisiologi, che infatti, nell’ultimo secolo, li hanno caratterizzati in dettaglio: ora sappiamo che all’attività del sonno sottostà una specifica attività neuronale chiamata slow oscillation. La slow oscillation, che è costituita da (ancora una volta) un’alternanza tra periodi di attività e periodi di iperpolarizzazione e silenzio neuronale (OFF-periods), è la modalità base di attivazione del cervello dormiente. Questa alternanza è dovuta alla tendenza dei neuroni surante lo stato di sonno, di passare ad un periodo silente dopo un’attivazione iniziale, una tendenza a cui viene dato il nome di bistabilità neuronale. Molti studi hanno dimostrato come la bistabilità neuronale tipica del sonno ed i relativi OFF-periods, possano accadere anche durante la veglia in particolari condizioni patologiche, nelle transizioni del sonno e durante le deprivazioni di sonno. Per questo motivo, se accettassimo che la bistabilità neuronale e gli OFF-periods rappresentino una caratteristica fondamentale del sonno, allora dovremmo ammettere che stiamo assistendo ad un cambio di paradigma: da una prospettiva neurofisiologica il sonno può intrudere nella veglia. In questa tesi ho analizzato i nuovi -fluidi- confini tra sonno e veglia e le possibili implicazioni di questi nel problema della persistenza personale attraverso il tempo. Inoltre, ho studiato le implicazioni cliniche dell’intrusione di sonno nella veglia in pazienti con lesioni cerebrali focali di natura ischemica. In particolare, i miei obiettivi sono stati: 1) Dimostrare come la bistabilità neuronale possa essere responsabile della perdita di funzione nei pazienti affetti da ischemia cerebrale e come questo potrebbe avere implicazioni nello studio della patofisiologia dell’ischemia cerebrale e nella sua terapia; 2) Stabilire le basi per un modello di sonno locale presente nella vita di tutti i giorni: la sensazione di sonnolenza. Infatti, essa potrebbe riflettere la presenza di porzioni di corteccia in stato di sonno, ma durante lo stato di veglia; 3) Difendere il criterio biologico di identità, che troverebbe nell’attività cerebrale la continuità necessaria al mantenimento della nostra identità nel tempo.
Sleep and wakefulness are considered two mutually exclusive states. The alternation between those two states seems to be a defining characteristic of our life, a ubiquitous phenomenon demonstrated in every animal species investigated so far. However, during the last decade, advances in neurophysiology have blurred the boundaries between those states. The mechanisms of sleep have always intrigued neurophysiologists and great advances have been made over the last century in understanding them: we now know that the defining characteristic underlying sleep activity is a specific pattern of neuronal activity, namely the slow oscillation. The slow oscillation, which is characterized by the periodic alternation between periods of activity (ON-periods) and periods of hyperpolarization and neuronal silence (OFF-periods) is the default mode of activity of the sleeping cortex. This alternation is due to the tendency of neurons to fall into a silent period after an initial activation; such tendency is known as “bistability”. There is accumulating evidence that sleep-like bistability, and the ensuing OFF-periods, may occur locally in the awake human brain in some pathological conditions, in sleep transition, as well as after sleep deprivation. Therefore, to the extent that bistability and OFF periods represents the basic neuronal features of sleep, a paradigm shift is in place: from a neurophysiological perspective sleep can intrude into wakefulness. In this thesis, I explore the fluid boundaries between sleep and wakefulness and investigate their possible implications on the problem of personal persistence over time. Moreover, I study the clinical implications of the intrusion of sleep into wakefulness in patients with focal brain injury due to stroke. Specifically, I aim to: 1) show how the sleep-like bistability can be responsible for the loss of function in stroke patients. This may have implications for understanding the pathophysiology of stroke and helping to foster recovery; 2) establish the basis for a model of local sleep that might be present in the everyday life, id est the sensation of sleepiness. Indeed, sleepiness could reflect islands of sleep during wakefulness; 3) advocate the biological criterion of identity, in which the continuity necessary for maintaining ourselves over time could be represented by never resting activity in the brain.

Analysons-nous le monde de façon continue ou bien selon une séquence d'évènements, un peu comme des instantanés pris par une caméra vidéo ? C'est la question qui a motivé ma thèse dans un premier temps. De précédentes expériences ont démontré que l'information visuelle était échantillonnée de façon périodique par l'attention, et que ce traitement était supporté par des oscillations de l'activité EEG. Dans le 1er papier, en utilisant la TMS, nous avons pu établir pour la première fois une relation causale entre la phase des oscillations spontanées, l'excitabilité cérébrale et la perception visuelle. Dans une autre série d'expériences, nous nous sommes demandé quel était le comportement spatio-temporel de l'attention au cours de tâches de recherche visuelle. A l'aide de diverses expériences (papiers 2 à 4) et de différentes techniques (TMS, EEG, psychophysique), nous avons pu établir des arguments convaincants en faveur d'un échantillonnage périodique de l'information visuelle par l'attention. De plus, dans le 5ème papier, nous avons pu clarifier une question hautement débattue concernant les tâches de recherche visuelle en éliminant la possibilité d'un traitement en parallèle de l'intégralité des stimuli présents à l'écran, suggérant un traitement séquentiel des différents stimuli au cours de la recherche. Ce travail de thèse a permis d'apporter des arguments forts en faveur d'un traitement périodique, voire séquentiel, de l'information visuelle par l'attention
Do we experience the world continuously or as a discrete sequence of events, like samples of a video camera? This is the first question motivating my PhD work. Previous experiments have shown that visual information may be sampled periodically by attention, this processing being supported by oscillations in the EEG brain activity. In paper 1, using TMS, we were able to establish for the first time a causal relation between the phase of ongoing oscillations, brain excitation and visual perception. In another series of experiments, we explored the spatio-temporal behaviour of attention during visual search tasks. Using various experiments (papers 2 to 4) and various techniques (TMS, EEG, psychophysics), we brought convincing and converging evidence in favour of a periodic sampling of visual information by attention. Moreover, in paper 5, we were able to clarify an age-old debate concerning visual search tasks by ruling out the possibility that attention is distributed in parallel over all stimuli in the search array, suggesting a sequential processing of the different stimuli during the search. Overall, this PhD work gives strong arguments in favour of a periodic, and perhaps sequential, processing of visual information by attention

This thesis aimed to explore the relationship between emotional processing and the sensorimotor system, mainly focusing on one information source derived from emotional body language (EBL). We investigated such a relationship in four different experiments and through several methodologies ranging from behavioral to neurophysiological techniques, by means of transcranial magnetic stimulation (TMS) and high-density electroencephalography (hdEEG), in healthy subjects (experiments 1, 2 and 3) and patients affected by Parkinson’s Disease (PD) (experiment 4). In the first experiment, whose aims were to explore the ability to process, discriminate and recognize emotional information carried by body language and to test motor response through response times (RTs) to emotional stimuli (i.e., EBL and IAPS), we found that fearful EBL is rapidly recognized and processed, probably because of a rapid and instinctual activation of several brain structures involved in defensive reactions. In the second experiment we investigated the effects of emotion processing (i.e., Fear, Happy and Neutral) on the sensorimotor system through a TMS protocol assessing short-latency afferent inhibition (SAI) at two timepoints (i.e., 120 and 300 ms). Our results showed that sensorimotor inhibition in the first 120 ms after stimulus onset is increased during processing of fearful emotional stimuli, reflecting the fact that automatic processing of threatening information can modulate attentional resources and cholinergic activity. In the third experiment, were a protocol involving hdEEG and a source localization workflow was implemented in the study of event-related potentials (ERPs) and mu-alpha and beta-bands rhythms during EBL processing, we confirmed what observed in the second experiment by showing that during processing of fearful body expressions there was an increased activity in the β frequency band in the somatosensory cortex which in turn may be one of the factors responsible for reducing the activation of motor related areas and, hence, increase sensorimotor inhibition. Lastly, in the fourth experiment we partly replicated the experimental design of the first experiment but in patients with Parkinson’s disease and using not only emotional body language stimuli and emotional scenes, but also emotional facial expressions. Our results showed that motor responses in PD patients are speeded when observing a potential threat, for both the embodied set of stimuli (EBL and facial expressions). We discussed this finding in relation to the “Kinesia paradoxa” phenomenon, defined as “the sudden transient ability of a patient with PD to perform a task he or she was previously unable to perform”.

Language production is one of the most complex cognitive – motor skills developed by homo sapiens throughout evolution to allow inter-personal and intra-personal communication (Indefrey and Levelt, 2000). A great number of cortical regions have adapted to support this high-speed combination of muscular and mental processes, in order to correctly generate the intended utterances in different contexts and situations. The neural organization of language processing is a thorny matter, that in the last decades has been investigated with a number of different methods ranging from functional imaging (fMRI, PET; see Gernsbacher & Kashack, 2003; Price, 2010; 2012 for reviews) neurophysiology (EEG, ERP, MEG, see Ganushchak et al., 2011 for a review), lesion studies (for a review see Turkeltaub et al., 2011) and non-invasive brain stimulation (such as transcranial magnetic stimulation, TMS, and transcranial direct current stimulation, tDCS, Devlin and Watkins, 2007; Monti et al., 2012 for reviews). Overall these studies have identified specific areas differently involved in language sub-processes (for a review see Price, 2012; Indefrey, 2011). As a new methodology to investigate the relationship between cortical areas and behavioural performance in cognitive tasks, including language, tDCS has been increasingly used in the last decade (Vallar & Bolognini, 2011). This technique relies on a sub-threshold polarization or de-polarization of neurons that leads to a modulation of cortical excitability and plasticity (Nitsche & Paulus, 2011). Due to its ease of application even in clinical settings, the potential of this tool in neuro-scientific investigation seems wide, but there is no precise knowledge of its mechanisms and effects on cognitive functions. The aim of the present study is to test tDCS effects on language production, to explore when this technique can be applied and to deeply investigate the mechanisms that lead to behavioural changes. In particular, since one of the classification criteria in aphasia is verbal fluency, in study 10 1 I investigated the effects of anodal tDCS on a verbal fluency task, aiming at developing a possible protocol to apply on clinical populations. To assess whether stimulation could modulate language production, healthy subjects performed a verbal fluency task both on phonemic and semantic cue immediately after real or sham stimulation. Since this requires the activity of a distributed network, including, among others, the left inferior frontal gyrus (LIFG), the left pre-motor cortex (LPMC), the left inferior and superior temporal gyri (LITG, LSTG) and the bilateral occipital-temporal sulci (Birn et al., 2010), and given that widespread effects of tDCS on functional networks need further clarification, in study 2 I investigated how electrical non-invasive brain stimulation affects cortical excitability by means of a TMS- EEG and tDCS combination, assessing how tDCS modulates cortical excitability and, accordingly, behavioural performance on verbal fluency. An open issue, indeed, is how stimulation enhances the activity of functional networks during task execution. Few recent studies addressed this question, but they generally rely on imaging data (Meinzer et al., 2012, 2013; Holland et al., 2011;). Hence, I tested how cortical excitability is modified after anodal tDCS applied over the LIFG in a functionally connected area, namely the LPMC (BA6) and in a region not involved in verbal fluency, the left superior parietal lobe (LSPL, BA7), and whether these changes could explain the effects of tDCS on task performance. Then, in study 3 and 4 and 5, I tested whether tDCS could be a useful tool to investigate language processes in healthy subjects. In particular, in study 3 and 4 I focused on semantic and phonological interference in picture naming tasks. The functional locus of the semantic interference (SI) effect, indeed, is still not clear (Finkbeiner and Caramazza, 2006; Schnur et al., 2006; 2009; Schnurr and Martin, 2012) and the role of the LIFG and LSTG in this effect is still under debate (Schnur et al., 2006; 2009). To test the different hypotheses underlying SI effect, I investigated the effects of anodal stimulation on the two aforementioned areas in a naming task in which 11 semantic context was manipulated (“blocked naming task”, Belke et al., 2005). Similarly, frontal and temporal regions seem to be involved in the phonological facilitation (PF) effect observed in naming (De Zubicaray et al., 2002; De Zubicaray and McMahon, 2009; Zhao et al., 2012; Damian & Bowers, 2009; Meyer and Schriefers, 1991; Scrhiefers et al., 1990). A picture word interference paradigm (PWI) was then administered after anodal stimulation of the LIFG and LSTG, and the effect of stimulation on PF was assessed. Finally, since proper name retrieval decreases with aging (Evrard et al., 2002), it would be of high interest to develop protocols improving this ability: this is the topic of study 5.

Le interazioni sociali richiedono che un agente sia in grado di selezionare ed elaborare informazioni ambientali rilevanti, che sia situato in un contesto complesso, e che interagisca con altri agenti, rispettando le opportunità e i vincoli di contesto. Riconoscere noi stessi e gli altri come agenti intenzionali è un passaggio cruciale per il processo generale di comprensione sociale e, in particolare, per la nostra capacità di percepire le intenzioni e gli scopi altrui. Tali competenze sociali sostengono il nostro sviluppo fisico, cognitivo e affettivo promuovendo interazioni adattive. Di conseguenza, una disfunzione di tali competenze può compromettere gravemente l’autonomia e la qualità di vita. Si ritiene che un sistema distribuito medi la percezione di agentività e degli stati mentali altrui, ma la struttura interna dei processi che costituiscono la nostra capacità di comprendere i nostri simili e di interagire adeguatamente è tuttora per buona parte sconosciuta. Il progetto ha come obiettivo indagare le fasi iniziali di tali processi e, in particolare, l’elaborazione precoce di cues sociali (social affordances) per la detezione di agentività e opportunità d’interazione in contesti sociali. È strutturato in tre studi principali: il primo mira a esplorare i correlati elettrofisiologici (ERPs e dati di source localization) dell’elaborazione di informazioni visive per la detezione di agentività in interazione; il secondo mira a indagare possibili marcatori (ERPs) del profilo delle competenze di comprensione sociale associate alla sindrome di Williams; il terzo ha testato, tramite TMS, il ruolo causale di rTPJ nel mediare l’elaborazione pre-riflessiva di agentività e intenzionalità nel comportamento osservato.
Social interactions require an agent to be able to select and process relevant environmental information, to be situated in a complex context and to interact with other agents, according to the opportunities and boundaries of that context. Sensing ourselves and detecting others as intentional agents is a crucial step for the overall social understanding process and, in particular, for our ability to perceive others’ intentions and goals. Those social skills foster our physical, cognitive and affective development by promoting adaptive interactions. Consequently, a dysfunction of such skills can seriously affect the autonomy and quality of life. A distributed system is thought to subserve the perception of agency and others’ mental states, but the internal structure of processes that constitute our ability to understand our similars and interact adequately is still largely unknown. This project aimed at investigating early stages of those processes and, in particular, the initial elaboration of social cues (social affordances) for the detection of agentivity and opportunities for interaction in social situations. It is structured in three main empirical studies: the first one aimed at looking electrophysiological correlates (ERPs and source localization data) of visual information processing for the detection of agency in interactions; the second one aimed at looking for possible markers (ERPs) of the uneven profile of basic WS social understanding; the third one tested the causal role of rTPJ in mediating pre-reflective processing of agency and intentionality from observed behaviour by means of TMS.

Le interazioni sociali richiedono che un agente sia in grado di selezionare ed elaborare informazioni ambientali rilevanti, che sia situato in un contesto complesso, e che interagisca con altri agenti, rispettando le opportunità e i vincoli di contesto. Riconoscere noi stessi e gli altri come agenti intenzionali è un passaggio cruciale per il processo generale di comprensione sociale e, in particolare, per la nostra capacità di percepire le intenzioni e gli scopi altrui. Tali competenze sociali sostengono il nostro sviluppo fisico, cognitivo e affettivo promuovendo interazioni adattive. Di conseguenza, una disfunzione di tali competenze può compromettere gravemente l’autonomia e la qualità di vita. Si ritiene che un sistema distribuito medi la percezione di agentività e degli stati mentali altrui, ma la struttura interna dei processi che costituiscono la nostra capacità di comprendere i nostri simili e di interagire adeguatamente è tuttora per buona parte sconosciuta. Il progetto ha come obiettivo indagare le fasi iniziali di tali processi e, in particolare, l’elaborazione precoce di cues sociali (social affordances) per la detezione di agentività e opportunità d’interazione in contesti sociali. È strutturato in tre studi principali: il primo mira a esplorare i correlati elettrofisiologici (ERPs e dati di source localization) dell’elaborazione di informazioni visive per la detezione di agentività in interazione; il secondo mira a indagare possibili marcatori (ERPs) del profilo delle competenze di comprensione sociale associate alla sindrome di Williams; il terzo ha testato, tramite TMS, il ruolo causale di rTPJ nel mediare l’elaborazione pre-riflessiva di agentività e intenzionalità nel comportamento osservato.
Social interactions require an agent to be able to select and process relevant environmental information, to be situated in a complex context and to interact with other agents, according to the opportunities and boundaries of that context. Sensing ourselves and detecting others as intentional agents is a crucial step for the overall social understanding process and, in particular, for our ability to perceive others’ intentions and goals. Those social skills foster our physical, cognitive and affective development by promoting adaptive interactions. Consequently, a dysfunction of such skills can seriously affect the autonomy and quality of life. A distributed system is thought to subserve the perception of agency and others’ mental states, but the internal structure of processes that constitute our ability to understand our similars and interact adequately is still largely unknown. This project aimed at investigating early stages of those processes and, in particular, the initial elaboration of social cues (social affordances) for the detection of agentivity and opportunities for interaction in social situations. It is structured in three main empirical studies: the first one aimed at looking electrophysiological correlates (ERPs and source localization data) of visual information processing for the detection of agency in interactions; the second one aimed at looking for possible markers (ERPs) of the uneven profile of basic WS social understanding; the third one tested the causal role of rTPJ in mediating pre-reflective processing of agency and intentionality from observed behaviour by means of TMS.

The pathophysiology of psychiatric disorders such as schizophrenia is not fully understood due, in part, to the shortcomings of available neurophysiological techniques. Previous studies have shown that patients with schizophrenia have deficits in dorsolateral prefrontal cortex (DLPFC). In this regard, two major deficits were observed: impairments in gamma-aminobutyric-acid (GABA) neurotransmission and cortical gamma (30-50Hz) oscillations. Previous in vitro and animal studies have linked the modulation of gamma oscillations with GABAB receptor mediated inhibition. Objectives: The first objective was to examine the effect of GABAB receptor mediated inhibition on cortical oscillations in the motor cortex and DLPFC in healthy subjects by using the novel technique of transcranial magnetic stimulation (TMS) combined with electroencephalography (EEG) and through the TMS paradigm long interval cortical inhibition (LICI), which has been associated with GABAB receptor mediated inhibition. Second, to evaluate the psychometric properties of this neurophysiological paradigm, the validity and reliability of EEG indices of LICI were examined. Finally, the effect of LICI on cortical oscillations was examined in the DLPFC and motor cortex of patients with schizophrenia compared to healthy subjects and patients with bipolar disorder. Hypothesis: It was predicted that EEG measures of LICI would show validity and reliability, and it was hypothesized that patients with schizophrenia would show deficits in inhibition of gamma oscillations in DLPFC compared to healthy subjects and patients with bipolar disorder. Results: The first experiment showed that in healthy subjects LICI inhibited gamma oscillations in the DLPFC but not in the motor cortex. The second experiment demonstrated the validity and reliability of EEG indices of LICI were confirmed in healthy subjects. Finally, patients with schizophrenia had a selective deficit in inhibition of gamma oscillations in the DLPFC which appeared to be independent of illness duration or antipsychotic medication, and it was not observed in bipolar disorder. Conclusions: TMS combined with EEG allows for measuring modulatory effect of LICI on cortical oscillations. Inhibition of gamma oscillations in the DLPFC may be an essential neurophysiological process that may be impaired in schizophrenia. Future studies should ascertain the potential of gamma inhibition deficit as a biological marker for this illness.

Tese de mestrado integrado em Engenharia Biomédica e Biofísica, apresentada à Universidade de Lisboa, através da Faculdade de Ciências, 2015
Atualmente, no mundo das neurociências, a conectividade cerebral é um tema em destaque. Este conceito encontra-se dividido em conetividade estrutural (relações anatómicas entre estruturas cerebrais), conectividade funcional (dependências estatísticas entre estruturas cerebrais) e conectividade efetiva (relações de causalidade entre estruturas cerebrais). Esta tese debaterá fundamentalmente sobre o último destes conceitos, tentando oferecer uma interpretação para o fluxo de informações entre as áreas do cérebro. Muitas técnicas podem ser utilizadas na sua análise, entre os quais a Causalidade de Granger (GC) ou a estimulação magnética transcraniana em combinação com eletroencefalografia (TMS-EEG). Por um lado, a GC permite uma interpretação das ligações diretas dentro e fora das mesmas áreas cerebrais, sendo uma abordagem explicativa sobre os dados, onde não é necessária nenhuma hipótese sobre o comportamento das relações causais. No entanto, os resultados de GC são muito sensíveis, uma vez que dependem de sinais não-estacionários e não colineares, aspetos bastante presentes em sinais de eletroencefalografia (EEG). Desta forma, a qualidade dos resultados de causalidade irá sempre depender da qualidade do pré-processamento do sinal original, onde se tenta ao máximo reduzir o seu efeito tentando não alterar os padrões de conectividade artificialmente. Por outro lado, TMS é uma técnica que permite que a estimulação do cérebro através da despolarização de certas populações de neurónios, criando uma "onda" de propagação ao longo do cérebro, de acordo com o local de estimulação e as suas ligações fisiológicas. Estes sinais são medidos através de sistemas de EEG, fazendo de TMS-EEG uma poderosa ferramenta nos estudos de conectividade efetiva, uma vez que permite criar estados independentes da intenção consciente da pessoa, garantindo um acompanhamento abrangente da sua propagação. A combinação destas duas ferramentas poderosas (GC e TMS-EEG) permitirá uma abordagem inovadora no desenvolvimento do conhecimento relativo à conectividade efetiva no cérebro. É com essa intenção que esta tese foi desenvolvida, tendo criado uma ferramenta computacional que permita medir e inferir padrões de conectividade efetiva através da combinação de TMS-EEG com GC. Sendo uma abordagem pioneira, a tese foi estruturada para que inicialmente se desenvolvessem garantias relativas a que uma redução nos efeitos ambíguos da GC, como a não-estacionaridade e não-colinearidade dos sinais de EEG, não afetasse a qualidade dos resultados de causalidade, ou que minimizasse a sua dependência dos métodos de pré-processamento. Portanto, este projeto visa, em primeiro lugar encontrar respostas na abordagem da GC, tanto para as suas limitações como para os seus parâmetros, permitindo que, posteriormente, houvesse uma otimização da análise de dados de TMS-EEG. Nesta primeira fase, os testes foram realizados com dados simulados. Só numa segunda fase é que, o objetivo principal a que esta tese se propunha, foi alcançado. Para isso foi criada uma toolbox em Matlab (Effective Connectivity test Toolbox - ECt) permitindo uma combinação compatível de GC com TMS-EEG. Esta projeto tentou validar esta toolbox para que se torne uma ferramenta futura para estudos de conectividade cerebral. Sendo um pouco mais específico, na primeira fase, foi encetada uma comparação entre os métodos de estimadores de conectividade do cérebro. A tradicional implementação de GC foi comparada com um método inovador de combinar a modelação fatorial com a GC (FM-GC), e com duas aplicações de transferência de entropia (onde dois métodos de estimadores de entropia foram utilizados – Binning Estimator e k-Nearest Neighbor). Esta abordagem mostrou que, perante os dados simulados criados, a GC se adaptou melhor tanto ao ruído implementado no sistema, comprovando ser o método com maior sensibilidade e especificidade. Provou-se também que para condições reais de EEG, nomeadamente número de pontos por trecho (512) e o número de ensaios (1 ensaio) a utilizar, GC verificou valores de falsos positivos menores comparativamente com os outros métodos. De seguida foram consideradas e testadas soluções que permitissem suavizar os efeitos da não-estacionariedade e colinearidade nos resultados da GC, tentando perceber novamente o desempenho deste método em dados simulados. Relativamente aos métodos de não-estacionaridade, o demean e o detrending foram implementados sendo que foi também analisada a capacidade de redução da presença de ‘raízes unitárias’, conduzindo à atenuação da não-estacionaridade, através do Augmented Dickey-Fuller (ADF). Este protocolo foi aplicado em dois tipos dados simulados – estacionários e não-estacionários. Foi também analisada a possibilidade de tornar os modelos autorregressivos (MVAR) mais estáveis através da conjugação de vários ensaios. Relativo ao primeiro teste pouca diferença foi verificada, no entanto conclui-se que inclusão desses métodos (demeaning & detrending) deveria ser introduzida na pipeline de pré-processamento. Relativamente à segunda etapa provou-se a eficácia de um aglomerar de ensaios sendo que o valor que otimizava essa estabilidade era de 5 ensaios. Por fim, testaram-se e debateram-se métodos que reduzissem a colinearidade e o overfitting do modelo. Relativo ao problema de colinearidade foi debatido, com base nas referências bibliográficas, que a implementação de uma solução para o problema inverso (encontrar matematicamente as fontes de sinal de EEG) seria necessário para remover essa ambiguidade. Sendo que a escolha recaiu sobre a análise de componentes independentes (ICA), assumindo que cada componente independente assegura o comportamento de uma fonte de sinal elétrico no cérebro. Relativamente ao overfitting verificou-se apenas, num sinal simulado de ERP, que com o acrescer de sensores (variáveis) existe um aumentar de parâmetros que traduzem o overfitting como, os coeficientes de correlação (relação diretamente proporcional) ou as intensidades máximas de GC (que foram otimizadas para um número de sensores entre 15 e 20). Na segunda parte do projeto, e de forma a responder ao objetivo principal, foi realizada uma experiência de TMS-EEG, que por um lado permitisse garantir dados realistas e provenientes dessa modalidade, como por outro que permitisse validar as mais-valias da toolbox ECt. Nesse âmbito, foram realizadas, a seis sujeitos, duas condições, uma estimulação real de TMS e uma inovadora estimulação sham, que foram repetidas cercas de 200 vezes. Em ambos os casos, o foco esteve no período de repouso (rs) entre os pulsos de TMS. Nestas condições, foi possível tentar validar a eficácia da ECt, pois no período de repouso após um pulso de TMS o comportamento de ambas as condições era suposto de ser idêntico, não se esperando mudanças ao nível da conectividade. Isto levou a formular a hipótese de que quando comparando as duas condições, estatisticamente os resultados não seriam significativamente diferentes. Colocada a hipótese, estruturou-se a toolbox em três pontos. O primeiro recaiu sobre os métodos de pré-processamento. Este abordou aspetos relativos ao tratamento dos dados de EEG recolhidos, onde se procedia a uma correção da baseline, procedia a uma redução da frequência de amostragem, se concatenavam os dados de todos os sujeitos para que posteriormente a abordagem de ICA fosse mais coerente. No segundo ponto debateu-se a aplicação do método da GC, onde se realizaram procedimentos como: estimar a ordem do modelo, estimar o modelo MVAR e, posteriormente, calcular os índices de GC. O último ponto incidiu sobre uma abordagem estatística inovadora de três níveis de análise. O primeiro pretendia validar os resultados de GC dentro de um sujeito e de uma das condições, através de uma análise estatística dos resultados de GC contra um surrogate (teste inverso de granger - RGT). O segundo nível pretendia comparar os resultados de GC entre condições dentro de um sujeito (Maximum permutation statistics). Por último, o terceiro nível tinha como objetivo comparar os valores de GC entre condições e sujeitos (t-test paralelo). Os resultados permitiram, numa primeira fase, verificar que os métodos de pré-processamento permitiram a redução de conectividade espúria, uma vez que 10 em 12 (2 condições vezes 6 sujeitos) dos conjuntos de dados preservaram mais de 60% dos ensaios sobrevivendo às restrições impostas pelo modelo. Considerando os resultados estatísticos obtidos, e tendo em consideração a falta de sujeitos (apenas 6 indivíduos), eles parecem ser promissores já que não existe uma expressão significativa nas matrizes de causalidade quando comparadas no segundo nível de análise estatística (Comparação de um sujeito entre condições), onde apenas 9% das ligações possíveis foram estatisticamente significativas. Relativamente ao último nível de análise os resultados não mostram qualquer inferência significativa entre variáveis, muito provavelmente devido ao fraco poder estatístico (apenas 6 sujeitos) do procedimento realizado. Para concluir, todos os aspetos considerados e discutidos nesta tese, relativos tanto a esta teoria como a esta toolbox podem e deveram ser consideradas como um primeiro passo, visto que este projeto visou criar uma base para o estudo da conectividade efetiva em protocolos de TMS. No futuro pode permitir abrir a porta à compreensão da complexidade da estrutura de causalidade e dinâmica do sistema cerebral.
Brain connectivity is a ‘hot’ topic these days in neuroscience. One of its branches is the effective connectivity, which intends to offer an interpretation for the information flow across brain areas. Many techniques can be used, between which Granger Causality (GC) and transcranial magnetic stimulation in combination with electroencephalography (TMS-EEG) have a prominent position. On one hand GC allows an interpretation of direct connections among brain areas, being an explanatory approach over the data, where no assumptions regarding the behavior of the causal relations are needed. However, several issues affect the results of GC, since they have to be contained over the restrictions of the model (i.e. non-stationarity and colinearity) and they are highly affected by spurious causality making the statistical reliability tenuous. On the other hand, TMS is a brain stimulation technique that allows the depolarization of populations of neurons, creating a ‘wave’ of propagation over the brain, according to the place of stimulation and its physiological connections. When these waves are measured with an EEG system, a combination of TMS-EEG is made. Such technique can become a powerful tool in connectivity studies. Uniting these two powerful tools (GC and TMS-EEG) should allow a new and innovative approach to measure effective connectivity in the brain. However, as yet, no study was made coupling these two methods. Therefore this project aims firstly to find answers in the GC approach for both its limitations and its parameters, allowing for the optimizations of the further TMS-EEG data analysis with GC. Secondly, and as a major goal, to create a Matlab toolbox (Effective Connectivity test - ECt), that allows the compatible combination of GC with TMS-EEG, making possible to use it as a future tool for brain connectivity studies. In the first phase, a comparison between methods of brain connectivity estimators was made (GC was compared with: Factor Modelling combined with GC, and with Transfer Entropy), showing that GC outperformed others. It was also taken into account and tested solutions for the non-stationarity and colinearity of the data over simulated data. Such procedure allowed to select specific GC parameters such as, number of ensemble trials, data length and number of variables. In a second part of the project, a TMS-EEG experiment was performed. Two conditions were recorded, a real TMS stimulation and a groundbreaking sham stimulation. The focus was on the resting state (rs) period in between the TMS pulses, because in a single pulse stimulation on the rs no changes in terms of connectivity were expected. Thus, this procedure allowed to validate the toolbox on recorded data by comparing such two conditions. An innovative pre-processing and statistical approach on the GC was implemented and validated, allowing the reduction of spurious connectivity. Considering the results, having in considerations the lack of subjects (only 6 subjects), they look promising since no big effect is seen (less than 9% of connections are significant) over the statistical analysis. All things considered, the toolbox, techniques here discussed and its premises, can be considered as a first step into measuring effective connectivity with a coupling of two techniques such as TMS-EEG and GC. In the future this might lead to a better understanding of the structure complexity and system dynamics of the brain.

La corteccia parietale posterior destra (rPPC) è strettamente coinvolta nelle funzioni visuo-spaziali, come rivelato dagli studi sui pazienti affetti da neglect (Vallar, 1998) e da quelli con TMS (Fierro et al., 2000; Bjoertomt et al., 2002; Ellison et al., 2004; Fierro et al., 2006; Ricci et al., 2012). All’interno di questa cornice teorica, uno degli strumenti più utilizzati è senza dubbio il compito di Landmark (LT, Milner et al., 1992, 1993), un test in cui viene richiesto di giudicare la lunghezza di linee prebisecate e i cui correlati neurali sono ben conosciuti (Fink et al., 2000, 2001; Ҫiҫek et al., 2009). Altro punto di forza di tale strumento è la sua abilità nel distingure errori di natura percettiva da quelli di risposta (Bisiach et al., 1998). Partendo da questo quadro teorico, tramite l’utilizzo combinato di EEG e TMS, il nostro obiettivo è quello di indagare gli effetti comportamentali (la modulazione del bias percettivo, PB) e neurofisiologici (il cambiamento dell’attività neurale) indotti dalla stimolazione a singolo impulso della rPPC tramite TMS. L’esperimento si componeva dei seguenti passaggi: (i) procedura di hunting (Salatino et al., 2014), in cui dieci singoli impulsi venivano rilasciati per ognuno dei 9 punti di una griglia centrata su P6, mentre il soggetto era sottoposto a LT; (ii) somministrazione del LT durante la registrazione dell’EEG con (TMS ON) e senza (TMS OFF) stimolazione dell’hotspot parietale. In entrambe le condizioni veniva fatto uso di linee simmetricamente e assimetricamente bisecate. I partecipanti sono stati suddivisi in 3 gruppi a seconda della modulazione del PB nella condizione TMS ON rispetto a quella di TMS OFF: il gruppo Neglect-like bias (n=16, PB TMS ON> PB TMS OFF), il gruppo Pseudoneglect-like bias (n=14, PB TMS ON< PB TMS OFF), e il gruppo No Bias (n=14, PB TMS ON= PB TMS OFF). È stata anche condotta un’analisi spazio-temporale sulla differenza tra linee assimmetricamente vs simmetricamente bisecate per la condizione TMS ON e TMS OFF, rispettivamente per ogni gruppo (Groppe et al., 2011a & 2011b). In uno stadio iniziale del processing delle informazioni abbiamo trovato un effetto significativo negli elettrodi O2 e P8 nella condizione TMS OFF in due gruppi (Pseudoneglect e No Bias), che si è rivelato ancora presente, negli stessi siti, nella condizione TMS ON solo per il gruppo No Bias. Successivamente un effetto d’interazione significativo della condizione TMS per il tipo di stimolo è stato riscontrato solo per il gruppo No Bias negli elettrodi dell’emisfero sinistro. Infine, tra 200 e 430 ms, in tutti i gruppi, le onde derivanti dell’analisi sulla differenza tra i tipi di stimoli sono risultate significative in quasi tutti gli elettrodi. I dati fin’ora raccolti sembrano confermare il ruolo della rPPC nel giudizio sulla lunghezza di linee. È così possibile concludere che la TMS è stata in grado di indurre differenti tipi di modulazione del PB, anche se tale effetto non è risultato presente in un gruppo di partecipanti (No Bias). Questo effetto è probabilmente dovuto a differenze preesistenza tra i soggetti, come i nostri risultati ottenuti in uno stadio precoce di elaborazione, nella condizione senza stimolazione, sembrerebbero suggerire. Una possibile spiegazione è che l’effetto della TMS non sia determinato dalle proprietà dello stimolo stesso o dai parametri dello strumento, ma anche dallo stato della corteccia durante l’esecuzione del compito (Silvanto & Pascual-Leone, 2008). Tali risultati sembrerebbero inoltre sottolineare che i nostri gruppi siano differenti nel processing percettivo degli stimoli. Recentemente abbiamo iniziato a testare eventuali differenze tra i gruppi. Questo tipo di analisi permetterebbe di chiarire se, a livello neurale, il gruppo No Bias è significativamente diverso da quello Neglect e/o Pseudineglect e le preesitenti diversità riscontrate tra i nostri partecipanti.
The right posterior parietal cortex (rPPC) is involved in visuo-spatial processing, as neglect patients (Vallar, 1998) and TMS studies revealed (Fierro et al., 2000; Bjoertomt et al., 2002; Ellison et al., 2004; Fierro et al., 2006; Ricci et al., 2012). Within this framework, one of the most frequently used research tasks is the Landmark Task (LT, Milner et al., 1992, 1993), a line bisection judgments task whose neural correlates are well known (Fink et al., 2000, 2001; Ҫiҫek et al., 2009). Remarkably, it affords to disentangle perceptual and response biases (Bisiach et al., 1998). Given this background, by combining EEG and TMS, we want to investigate the behavioral (i.e. modulation of perceptual bias, PB) and neurophysiological (i.e. brain activity changes) effects of single pulse TMS over rPPC. The experiment followed the subsequent steps: (i) hunting procedure (Salatino et al., 2014), delivering ten single pulses for each of the 9 points of a grid centrally located over P6, while the subject was performing the LT; (ii) administration of the LT while recording EEG with (TMS ON) and without (TMS OFF) stimulation of the parietal hotspot. In both conditions, symmetrically and asymmetrically bisected lines were used. Participants were divided in three different groups depending on the modulation of the PB on the TMS ON condition as compared with the TMS OFF condition: the Neglect-like bias group (n=16, PB TMS ON> PB TMS OFF), the Pseudoneglect-like bias group (n=14, PB TMS ON< PB TMS OFF), and the No Bias group (n=14, PB TMS ON= PB TMS OFF). We also performed a spatiotemporal analysis on the difference between asymmetrical vs symmetrical lines for the TMS ON and the TMS OFF conditions, separately on each groups (Groppe et al., 2011a & 2011b). In an early stage of processing we have found a significant effects in O2 and P8 electrodes in the TMS OFF condition in two groups (Pseudoneglect and No Bias), that was still present, in the same sites, in the TMS ON condition only for the No Bias group. Later in time there was a significant interaction effect of the TMS condition on the type of stimuli for only the No Bias group in the electrodes of the left hemisphere. Finally, between 200 and 430 ms, in all the groups, the difference waves were significant in almost all electrodes. The present data thus show that rPPC is involved in magnitude estimation of line length. Generally we could conclude that the TMS induces different type of modulation of PB. Indeed TMS could not modulate the PB in a group of participants (No bias), probably due to preexisting differences between participants, as our results in the early time window in the TMS OFF condition would suggest. One possibility is that the effects of the TMS are determined not only by the properties of the stimulus or by the TMS itself, but also by the state of the cortex during the task execution (Silvanto & Pascual-Leone, 2008). These results seem to suggest that our groups are different in the perceptual processing of the stimuli. Recently we are testing differences between groups. These would help us to clarify if, at a neural level, the no bias group is significantly different from the neglect and pseudoneglect like bias group. We are also trying to better understand the pre-existing difference found in our participants.

I sostanziali progressi tecnologici occorsi nell’arco dell’ultimo secolo, a partire dalla osservazione dell’attività elettrica cerebrale (EEG) da parte di Berger negli anni ‘20 del secolo scorso ai recenti studi di optogentica (Adamantidis et al., 2013), hanno fatto luce su parte della fisiologia del sonno, in maniera impensabile fino alla introduzione di queste metodiche. Il sonno quindi è passato dall’essere considerato uno stato di assoluta inerzia, paragonabile alla morte, ad uno stato di attività cerebrale proteiforme, passando da stadi di ipersincronizzazione a stadi di attività cerebrale simili alla veglia, mantenendo sempre un certo grado di reattività a stimoli esterni perturbanti (Terzano and Parrino, 2000). Il sonno, o la sua mancanza, sono inoltre a loro volta in grado di modulare funzioni cerebrale: è partica clinica comune la privazione di sonno come tecnica di facilitazione per la comparsa di anomalie epilettiche. Questi fenomeni sono stati studiati dal punto di vista patofisiologico con l’osservazione di un abbassamento della soglia di eccitabilità neuronale nell’animale deprivato di sonno (Cohen and Dement, 1965) ed una facilitazione al kindling (Shouse, 1988) possibilmente dovuto ad uno sbilanciamento tra neurotrasmettitori eccitatori ed inibitori (Naitoh and Dement, 1974). Il limite di questo tipo di studi è però la loro invasività, così come lo era stato per gli studi pionieristici sulla fisiologia del sonno di Bremer (1935, 1936) e Moruzzi e Magoun (1949): la crescente consapevolezza delle problematiche etiche legate a questo tipo di esperimenti, unito alla loro non applicabilità sull’uomo, ha spinto verso la ricerca di nuove metodiche non invasive ed applicabili in vivo. All’inizio degli anni ’80 è stata introdotta la risonanza magnetica cerebrale (RMN) che ha permesso, oltre alla più dettagliata definizione anatomica, lo studio di attivazione di aree cerebrali implicate in task specifici (risonanza magnetica cerebrale funzionale) grazie alla implementazione di specifiche sequenze che identificano la variazione di segnale determinata dalla quantità di desossiemoglobina che si forma in quella area cerebrale e che presenta proprietà paramagnetiche, rispetto ad una condizione di riposo. Si ottiene così il cosiddetto segnale BOLD (blood oxygenation level dependent), che permette di identificare le aree cerebrale a maggior consumo metabolico all’interno della finestra di tempo studiata, e che quindi sono correlabili al task eseguito. L’osservazione casuale di una attivazione BOLD task-indipendente ha portato alla formulazione del concetto di default mode network (DMN), una serie di aree cerebrali costanti inter-individuo che sembrano attivarsi una condizione di veglia rilassata (Raichle et al., 2001), e che, sorprendentemente, persistono sia pur modulate anche durante il sonno. I diversi network che sono presenti in questa condizione di riposo sono stati successivamente classificati grazie alla implementazione di modelli matematici di calcolo in grado di scomporre il segnale nelle sue costituenti: si è così arrivati alla definizione dei resting state networks (RSNs) (Rosazza and Minati, 2011), ognuno dei quali pare relato ad una specifica funzione fisiologica. Un ulteriore sviluppo tecnologico, introdotto circa 15 anni fa, è stata l’entrata in commercio di sistemi EEG compatibili con campi magnetici, siano questi in RMN, permettendo così la registrazione della attività cerebrale all’interno dello scanner, che durante stimolazioni magnetiche. La stimolazione magnetica transcranica (TMS), sfruttando il principio della induzione elettromagnetica, è in grado di applicare uno stimolo elettrico agli strati superficiali della corteccia, ed elettrodi EEG compatibili (sistemi di co-registrazione EEG-TMS) permettono di studiare gli effetti diretti di uno stimolo TMS sulla attività cerebrale (Ilmoniemi et al., 1997) senza le possibili interferenze delle vie discendenti, che contribuiscono a costituire il potenziale evocato motorio (Groppa et al., 2012). La modulazione che la TMS induce a livello della corteccia cerebrale può essere anche valutata in base alla modificazione dei ritmi cerebrali (Thut and Miniussi, 2009), che reagiscono in maniera differente a seconda del paradigma somministrato (Manganotti and Del Felice, 2012) o delle frequenze cerebrali intrinseche o della frequenza di stimolo (Thut et al., 2012). Infine, nell’ultimo decennio sono state realizzati e commercializzati sistemi EEG ad alta densità di elettrodi (256). Questa elevata risoluzione spaziale ha riportato in auge un modello di analisi matematica, la cosiddetta ricostruzione inversa della sorgente, tramite la quale si cerca di ricostruire il generatore profondo dell’attività elettrica registrata sullo scalpo (Fender, 1987; Brunet et al., 2011). Questa metodica, denominata electrical source imaging (ESI), sfruttando l’elevata risoluzione della cuffia a 256 canali e la possibilità di proiettare il dato sulla RMN del paziente, è quindi stata applicata all’identificazione della sorgente di grafoelementi epilettici (Scherg and Von Cramon, 1985, Liu et al., 1998, Babiloni et al., 2003, Michel et al., 2004), mentre sono ancora scarsi in letteratura dati sul sonno (Siniatchkin et al., 2010). Lo scopo di questa tesi è la discussione delle possibili applicazioni di queste tecnologie allo studio di quesiti ancora aperti nella fisiologia e patofisiologia del sonno. Un primo approccio è stato lo studio della modulazione della reattività corticale in soggetti sani e pazienti epilettici in sonno e privazione di sonno con co-registrazioni EEG-TMS, valutando sia le componenti evocate lente che le variazione dei ritmi indotte dallo stimolo. Un secondo set di esperimenti ha indagato tramite risonanza magnetica funzionale le attivazioni cerebrali in sonno durante una stimolazione elettrica. Infine le potenzialità dell’electrical source imaging sono state applicate per chiarire quale sia il numero e la localizzazione dei generatori corticali dei fusi del sonno in volontari sani ed in pazienti epilettici, e la correlazione delle sorgenti delle figure del sonno con il focus epilettogeno in questa ultima popolazione. I diversi aspetti neurofisiologici che queste tecniche indagano offrono una prospettiva sfaccettata dei fenomeni del sonno. La traslazione di questi paradigmi di studio a stati di ridotta coscienza dovrebbe essere una delle future prospettive di ricerca.
As has been stated, we have gained more knowledge on sleep physiology in the last 60 years than in the previous 6000 (Hobson, 1989). This holds true thanks to the massive advances technologies have provided in the past century, ranging from the introduction of electroencephalogram (EEG) in the twenties of the last century by Berger to the latest optogenetic approaches (Adamantidis et al., 2013). The major change put forward by this more detailed understanding of sleep function and functioning has been the transition of sleep as a state of absolute inertia, paralleled to death by almost all the ancient literature, to a reactive state of the brain: during sleep, cerebral activity presents its most diverse expressions, from the bold slow waves sleep of the deep stages to the wake like activity of REM associated with muscular atonia, and is able to differently react to external perturbations with rapid frequency shifts (Terzano and Parrino, 2000). Moreover, the modulations that sleep and sleep deprivation exert have been postulated deriving both from plain clinical observations, i.e. in epilepsy, and from animal studies. Sleep deprivation is the best method for provoking EEG epileptiform abnormalities and seizures (Bennett, 1963; Pratt et al., 1968; Jovanovic, 1991; King et al., 1998) in most types of epilepsy (Dinner, 2002), and many epileptic syndromes, such as the generalized idiopathic epilepsies (IGE), are prone to circadian fluctuations related to the sleep-wake cycle - with seizures gathering mostly early in the morning or at awakening (Niedermeyer et al., 1985). The mechanisms underlying the activation of paroxysmal activity remain to be elucidated. The activation of epileptic patterns has been attributed to drowsiness and sleep (Pratt et al., 1968), while sleep deprivation has been shown to have a specific activating effect on patients who remain awake during recording (Naitoh and Dement, 1974). In animals, sleep deprivation results in a lowering of the threshold for electroshock convulsions (Cohen and Dement, 1965) and kindling (Shouse, 1988) due to a shift in the balance between excitatory and inhibitory neurotransmitters (Naitoh and Dement, 1974). But while animal studies deploy invasive techniques, as did the pioneer physiology studies by (Bremer 1935 and 1936; Moruzzi and Magoun, 1949) that allowed the definition of cerebral and truncal structures involved in sleep building-up and maintenance and their neurotransmitters, growing concerns about in vivo animal studies have pushed towards other research methods, that moreover could be applied to the human being too. Indeed, one of the major limitations in the field of sleep research up to the last decades was determined by the only available technique applicable in humans - electroencephalogram. Since the eighties of the last century, a series of technological advances introduced in clinical practice Magnetic Resonance Imaging (MRI). MRI permits not only a more detailed visualization of brain structures than those of previous neuroimaging, such as computed tomography (CT) scanning, but also, due to the implementation of new acquisition sequences and analysis procedures, the identification of blood oxygenation level dependent (BOLD) activations. The latter consists of a cerebral area in which any sort of metabolic process is going on, in a frame time of a few seconds, and is generally related to areas active due to a given task. The serendipitous observation that persistent BOLD activated areas are present also in the idling brain led to the proposal of the concept of a default mode network (DMN), that is, a series of possibly interconnected cerebral regions that switch on in the very moment any brain engagement is supposed to switch off (Raichle et al., 2001). The persistence of an analogous pattern also during sleep led to the hypothesis of this network to be the neural substrate of mentation and perhaps consciousness. Further improvements in the mathematical models that support BOLD signal analysis were subsequently able to disentangle the various components of the this “resting brain activity”, generating an array of so called resting state networks (RSNs) (Rosazza and Minati, 2011) that encompass diverse physiological functions. A step further was possible with the introduction, almost 15 years ago, of MRI compatible EEG equipment that prevents the generation of oddy currents inside the electrode: the concomitant EEG registration with an MRI scan permits to relate a particular EEG activity with the underlying BOLD signal. The same magnetic field shielded electrodes were later on exploited in the contest of electro-magnetic fields generated through wires rolled into a coil, that were presented by Baker in 1985 as transcranial magnetic stimulation (TMS). TMS final effect is that of electrically stimulating the superficial layers of the cortex, and EEG-TMS co-registration offered the chance to investigate the direct effect of a pulse on the cortex (Ilmoniemi et al., 1997) by removing possible interferences from the descending motor pathways, that were intermixed in the standard parameter by which TMS alone is evaluated - the motor evoked potential (MEP) recorded from a muscle corresponded to the cortical activated area (Groppa et al., 2012). The perturbation TMS induces on the cerebral activity can also be studied as the modulation of EEG rhythms (Thut and Miniussi, 2009), that react differently depending on the stimulating paradigm (Manganotti and Del Felice, 2012) or on the intrinsic brain rhythm or stimulus frequency (Thut et al., 2012). The last technological innovation I am going to describe has been developed over the last decade: the introduction of high-density scalp EEGs (hdEEG), with up to 256 electrodes spread out over the scalp, the occiput and the cheeks of the subject, that offers a much higher spatial resolution than standard EEG caps. This high spatial resolution sampling has revived an older analysis method aimed at identifying via a mathematical approach called the inverse solution method the number, location and orientation of deep generators of scalp activity, the so called electrical source imaging (ESI) (Fender, 1987; Brunet et al., 2011). ESI involves numerous scalp electrodes, HdEEG , and realistic head models derived from structural MRI, and has so far mainly been applied to epileptic discharges (Scherg and Von Cramon, 1985, Liu et al., 1998, Babiloni et al., 2003, Michel et al., 2004), with only few reports in sleep (Siniatchkin et al., 2010). The aim of this dissertation thesis is to discuss the application of these technologies to the clarification of open issues in sleep physiology and pathophysiology. A first approach was to study the effects of sleep deprivation on cortical excitability through EEG-TMS co-registration experiments, both in healthy controls and in the frame of pathologically abnormal cortical excitability (i.e. epilepsy). A second set of experiments focused on fMRI data of subjects sleeping in the bore of the scanner during a concomitant external perturbation – an electrical stimulation at the wrist in the specific case. Finally, the potentiality of ESI has been applied to physiological sleep figures, in order to contribute to the open issue of their generators’ nature. A similar study design was also used in a population of focal epileptic patients, given the still actual debate over the relation of sleep figures and epileptic spikes. These techniques encompass different neurophysiological aspects providing a multiprospective view of sleep phenomena. The translation of such an approach to other states of reduced consciousness (i.e. vegetative or minimally conscious states) should be one of the future directions of research.

Questa tesi di dottorato si focalizza sullo studio sperimentale dei correlati neurali relativi alla percezione visiva e spaziale nell’uomo. Nello specifico, l’obiettivo di questo progetto è quello di identificare i marker elettrofisiologici della consapevolezza, sia in condizioni normali, sia quando la consapevolezza viene persa a causa di una lesione cerebrale reale o “virtuale”. In entrambi gli studi condotti è stato adottato un approccio causale multimodale, ovvero la coregistrazione dell’elettroencefalogramma durante la Stimolazione Magnetica Transcranica (EEG-TMS), che è in grado di fornire importanti informazioni rispetto ai correlati neurali della consapevolezza visiva e spaziale. Il primo studio condotto è centrato sula consapevolezza spaziale, e in particolare si focalizza sullo studio dei meccanismi alla base della sindrome del neglect. Il neglect viene definito come un deficit di consapevolezza in cui il paziente ignora gli stimoli che vengono presentati nel lato opposto rispetto a quello della lesione cerebrale. Uno dei modelli più influenti di spiegazione della disfunzione alla base di questa sindrome prende in considerazione il concetto di rivalità interemisferica, secondo cui l’emisfero intatto risulterebbe patologicamente iperattivato a causa della riduzione di inibizione precedentemente esercitata da parte dell’emisfero leso. L’obiettivo del presente studio è quello di testare questi modelli di rivalità interemisferica, investigando l’effetto che l’applicazione della TMS ripetitiva (rTMS) a bassa frequenza esercita sull’emisfero stimolato e su quello controlaterale nel processamento di stimoli visivi. In particolare, questo studio si pone l’obiettivo di indagare il contributo della corteccia parietale destra e sinistra in un funzionamento deficitario tipo neglect indotto attraverso l’applicazione della rTMS a bassa frequenza in un campione di soggetti sani. A quattordici volontari sani è stato richiesto di eseguire un test di bisezione di linee e un compito di detezione di stimoli visivi lateralizzati. Entrambi i compiti sono stati somministrati sia prima che dopo 30 minuti di rTMS a bassa frequenza (1 Hz) applicata sulla corteccia parietale posteriore destra. Il segnale EEG è stato registrato continuativamente per tutta la durata dell’esperimento. L’efficacia della rTMS nell’indurre sintomi tipo neglect è stata confermata dai risultati del compito di bisezione di linee, in cui i partecipanti hanno mostrato una performance simile a quella dei pazienti con neglect in questo test, ovvero con una deviazione verso destra del punto di bisezione dopo la rTMS. Considerando il compito di detezione, i risultati hanno mostrato come l’effetto della rTMS sia un rallentamento dei tempi di reazione sia per gli stimoli presentati a sinistra che per quelli presentati a destra e una riduzione dell’ampiezza della componente P200 registrata sui siti parietali sia sinistri che destri. I potenziali evocati dalla TMS (TEPs) registrati durante i 30 minuti di stimolazione hanno mostrato come la rTMS a bassa frequenza abbia indotto una riduzione dell’eccitabilità corticale sia della corteccia parietale destra (quella direttamente stimolata), sia delle aree omologhe controlaterali. Di conseguenza, i nostri risultati non hanno evidenziato una iperattivazione dell’emisfero sinistro conseguente all’inibizione dell’emisfero destro (come postulato dai modelli di rivalità interemisferica). Al contrario, l’inibizione della corteccia parietale destra ha indotto una propagazione dell’inibizione alle aree omologhe dell’emisfero sinistro. Il secondo studio condotto è focalizzato sulla consapevolezza visiva e, in particolare, si pone l’obiettivo di indagare i correlati neurali della percezione di fosfeni. La TMS a singolo impulso applicata sulla corteccia visiva primaria è in grado di indurre delle sensazioni visive, chiamate appunto fosfeni, che appaiono come brevi lampi di luce senza che effettivamente l’occhio venga stimolato da una luce. Recenti studi hanno dimostrato come la TMS possa produrre delle sensazioni visive non soltanto quando applicata sulle aree visive primarie, ma anche quando viene stimolata la corteccia parietale. Poiché le basi neurali coinvolte nella percezione dei fosfeni parietali sono tuttora sconosciute, rimane da capire se i fosfeni parietali siano generati direttamente da meccanismi locali o se invece la loro generazione dipenda dall’attivazione indiretta di altre aree visive. Al fine di caratterizzare i correlati elettrofisiologici della percezione dei fosfeni occipitali e parietali, abbiamo analizzato i TEPs in un campione di soggetti sani, confrontando i trials in cui i partecipanti riportavano di percepire un fosfene, con i trials in cui all’impulso della TMS non seguiva alcuna sensazione visiva. Quando viene stimolata la corteccia occipitale sinistra, la percezione dei fosfeni inizia a modulare i TEPs ad una latenza tardiva, mentre i fosfeni indotti dalla stimolazione della corteccia parietale sinistra iniziano a modulare i TEPs a delle latenze più precoci. Questa differenza tra fosfeni occipitali e parietali nell’andamento temporale dell’attivazione corticale potrebbe identificare un diverso meccanismo implicato nella loro generazione. L’effetto di latenza precoce osservato quando la TMS viene applicata sulla corteccia parietale potrebbe suggerire che i fosfeni parietali siano il risultato diretto dell’attivazione dell’area stimolata, e non la conseguenza di un’attivazione di tipo feedback della corteccia visiva primaria. Inoltre, abbiamo indagato i correlati elettrofisiologici della percezione di fosfeni parietali in una paziente emianopsica (SL) con una distruzione completa della corteccia visiva primaria sinistra. La percezione di fosfeni indotti dalla stimolazione della corteccia parietale ipsilaterale alla lesione ha mostrato lo stesso pattern di risultati rispetto alla stimolazione parietale del campione di soggetti sani, con una modulazione dei TEPs che emerge ad uno stadio di latenza precoce. Questi risultati supportano quindi la visione della corteccia parietale come un generatore indipendente di esperienze visive consce indotte dalla stimolazione magnetica.
This PhD thesis focuses on attempting to experimentally investigate the neural correlates of awareness related to visual and spatial perception in humans. Specifically, this project aimed at looking into the electrophysiological markers of awareness in normal conditions and when awareness is lost due to a real or a “virtual” lesion. In both the studies conducted we adopted a causal multimodal approach, namely Transcranial Magnetic Stimulation and EEG (TMS-EEG) co-registration, which can provide insights into the neural correlates of visual and spatial awareness. The first study focuses on spatial awareness and specifically on the investigation of the mechanisms underlying neglect syndrome. Neglect is defined as a disorder of consciousness in which patients fail to report, respond to, or orient to stimuli presented on the opposite side of the brain lesion. One of the most influential models to explain the dysfunction underlying this syndrome takes into account the concept of inter-hemispheric rivalry, which postulates a pathological hyperactivation of the unaffected hemisphere due to the reduced inhibitory influences from the lesioned hemisphere. The aim of the present study is to test these models analyzing the effect that low frequency repetitive TMS (rTMS) exerts on the stimulated and contralateral hemispheres in the processing of visual stimuli. Specifically we aim at assessing the contribution of left and right parietal cortices in an impaired neglect-like functioning induced by means of low frequency rTMS in healthy participants. Fourteen healthy volunteers performed a Line Bisection task and a simple detection task of unilateral checkerboards stimuli. Both tasks were performed either before and after 30 minutes of low frequency rTMS (1 Hz) over the right posterior parietal cortex. The EEG signal was continuously recorded throughout the experiment. The efficacy of rTMS in inducing neglect-like phenomena was confirmed by the results of the Line Bisection task where participants showed a rightward deviation after rTMS, a performance comparable to that of neglect patients. Detection task results showed that the effect of rTMS was a lengthening of reaction times for both left and right visual stimuli and a reduction of the amplitude of P200 component registered both on left and right parietal sites. TMS-evoked potentials recorded during 30 minutes of stimulation, showed that low frequency rTMS induced a reduction of cortical excitability both of the stimulated right parietal cortex and of the left contralateral homologous area. Therefore, our results did not show a hyperactivation of the left hemisphere due to the inhibition of the right hemisphere (as theorized by “rivalry models”). Conversely, the inhibition of the right parietal cortex induced a spreading of the inhibition to the homologous area of the left hemisphere. The second study focuses on visual awareness and specifically aimed at investigating the neural correlates of phosphene perception. Single-pulse TMS of the visual cortex is known to induce visual sensations, i.e. phosphenes, which appear as brief flashes of light without light actually entering the eyes. Recent studies have shown that TMS can produce visual sensations not only when it is applied over early visual areas but also when parietal cortex is stimulated. As the pivotal neural basis involved in the perception of parietal phosphenes still remain unknown, the main question is whether parietal phosphenes are generated directly by local mechanisms or emerge through indirect activation of other visual areas. To characterize the electrophysiological correlates of occipital and parietal phosphene perception we investigated TMS-evoked potentials in a sample of healthy participants by comparing trials in which a phosphene was perceived with trials in which no visual percept was reported. When the left occipital cortex was stimulated, phosphene perception started to affect TMS-evoked potentials at a late latency, whereas phosphenes elicited by left parietal cortex stimulation modulated TMS-evoked potentials at an earlier latency. This difference in the time-course of cortical activation between occipital and parietal phosphenes could underlie a different mechanism in their generation. The early latency of the phosphene effect observed when TMS was applied over the parietal cortex might suggest that parietal phosphenes should be considered as the direct result of the activation of the stimulated area, rather than the consequence of a feedback activation of the early visual cortex. Furthermore, we investigated electrophysiological correlates of parietal phosphene perception in a hemianopic patient (SL) who suffered from a complete destruction of the left primary visual cortex. Ipsilesional parietal phosphene perception in patient SL showed a similar pattern of results to that of parietal phosphene perception in healthy participants, starting to affect TMS-evoked potentials at an early stage of latency. This evidence might thus support the idea of parietal cortex as an independent generator of magnetically induced conscious visual experiences.

Contemporary research has been examining potential links existing among sensory, motor and attentional systems. Previous studies using TMS have shown that the abrupt onset of sounds can both capture attention and modulate motor cortex excitability, which may reflect the potential need for a behavioral response to the attended event. TMS, however, only quantifies motor cortex excitability immediately following the deliverance of a TMS pulse. Therefore, the temporal development of how the motor cortex is modulated by sounds can’t be quantified using TMS. Thus, the purpose of the present study is to use time frequency analysis of EEG to identify the time course of cortical mechanisms underlying increased motor cortex excitability after sound onset. Subjects sat in a sound attenuated booth with their hands outstretched at 45-degree angles while frequency modulated sounds were intermittently presented from a speaker either in the left and right hemispace. Our results indicated a transient reduction in EEG power from 18-24 Hz (300-600 ms latency) and then a long lasting increase in EEG power that began at ~800 ms and continued until at least 1.7 sec. The latency of EEG power changes was shorter for sounds presented from the right speaker at both time periods. When sounds were presented from the right speaker the contralateral hemisphere over motor regions also showed greater power increases after 800 ms relative to the ipsilateral hemisphere. In addition, power increases were greater in the left-handed subjects (8-12 Hz). Results showed that sounds increased EEG power at the time of a previously observed increase in motor cortex excitability. Findings also suggest an increased attentional salience to the right hemispace in neurologically normal subjects and asymmetrical hemispheric activations in right and left-handers.
acase@tulane.edu

We humans are visual creatures, constantly extracting information from the world around us. The source of our ability to understand the visual world is an intricate arrangement of multiple areas in our brains: the visual system. It enables us to recognize our friends and family in diverse conditions, to focus our attention on important aspects of a scene and performs invariant object categorization on multiple levels of abstraction. Vision has been in the focus of scientific interest for many decades and yet our knowledge of the cortical mechanisms involved is only limited. I here describe a series of experiments, in which we investigated how the visual system robustly and efficiently extracts meaning from the environment. In particular, I will focus on thee aspects of object recognition: sampling the environment, visual invariance, and categorization and plasticity. Starting with the selection of visual information, three eyetracking experiments are described in which we investigate the interplay of overt visual attention and object recognition. We show that overt visual attention and object recognition exert a bi-directional influence on each other. Whereas initial patterns of overt visual attention causally affect the outcome of the later recognition, briefly presented contextual information leads to substantial changes in the attentional sampling behavior, which can be best understood in terms of a shifting exploration-exploitation bias. Following this, we turn to visual processing within the system and ask how invariant object recognition is accomplished despite large variation in retinal input. As an exemplary case, we focus on changes introduced by rotations in depth. Using a variety of techniques, ranging from fMRI to TMS and EEG, we show that viewpoint symmetry, i.e. the selectivity to mirror-symmetric viewing angles, is a prevalent feature of visual processing across a wide range of higher-level visual regions. These findings jointly suggest that viewpoint-symmetry constitutes a key computational step in achieving full viewpoint invariance. On the next level of abstraction, we investigate how visual categories are represented at different levels of experience, from novice to expert. By combining training of novel visual categories with psychophysical measures, we demonstrate a change in the underlying type of category representation. Following this, we combine the training paradigm with electrophysiological measurements. In line with our behavioral results, these data reveal a spatiotemporal shift in category selectivity: from late and frontal to early occipitotemporal activity. These results suggest that novel and re-occurring categories rely on partially separate cortical networks, allowing the brain to balance robust and fast recognition with considerable flexibility and plasticity. The results of all experiments presented are unified by the concept of a system that has evolved efficient mechanisms for robust performance in a large variety of conditions. Using dynamic sampling strategies, computational shortcuts and a division of labor, the visual system is optimally equipped to support higher-level cognitive function in a complex and constantly changing environment.