The relationship between TMS-evoked activation of resting state networks and stroke impairment

Objectives: Impairment after stroke arises from direct anatomical damage and remote structural and functional abnormalities in large-scale resting state networks. A promising non-invasive technique to study brain connectivity with high temporal resolution is transcranial magnetic stimulation combined with electroencephalography (TMS-EEG). This methodology probes cortical reactivity through the registration of TMS-evoked cortico-cortical potentials (TEPs). In this study, we investigated patterns of cortical reactivity associated with motor impairment after stroke. Crucially, after stimulation of the motor cortex (M1), we assessed whole-brain TEPs and the evoked activation of the sensorimotor network (SMN). Materials and Methods: We recruited three chronic stroke patients with motor impairment. For each patient we collected demographic and clinical variables along with structural MRI scans. The first patient presented right hemiparesis after a left frontal lesion involving M1. The second patient showed left hemiplegia after a right corona radiata lesion. The last patient presented right hemiplegia after a left thalamic lesion. For all patients, TMS was applied over the ipsilesional and contralesional M1 while registering high-density EEG. All TMS/EEG data were pre-processed using a custom pipeline. Source reconstruction was used to estimate the distribution of activity which was mapped onto a functional parcellation atlas to quantify the activation of the SMN. The results were compared with normative data from 12 healthy controls. Results: Compared to healthy controls, ipsilesional TMS-evoked potentials revealed distinct activation profiles for patients with cortical and subcortical lesions, but no clear differences were observed in the patient with thalamic damage. Interestingly, SMN profiles of activation were altered in all patients compared to healthy controls. Cortical damage determined an excessive early activation of the SMN. Thalamic damage determined the loss of the physiological late activation of the SMN. Subcortical damage produced an intermediate pattern (i.e., prominent early activation, but reduced compared to perilesional stimulation). Discussion: Stimulation of M1 in stroke patients with motor impairment revealed distinct SMN activation patterns that varied according to lesion location. Perilesional stimulation detected a local prominent sleep-like cortical response, while thalamic damage was associated with the loss of late components possibly associated with sensorimotor integration. Evaluating TEPs at the network-level enabled a comprehensive assessment of areas potentially involved in post-stroke impairment, including spatially remote but functionally relevant regions. Conclusion: Our results provide insights into the neurophysiological significance of individual TEP components and large-scale network alterations following stroke. In this context, future studies could extend these findings beyond the sensorimotor system.