This project aims to tailor repetitive Transcranial Magnetic Stimulation (rTMS) target to the clinical profile of post-stroke aphasia, specifically focusing on lexicosemantic impairments, to enhance recovery. Functional Magnetic Resonance Imaging (fMRI) will provide critical insights into the effects of facilitatory versus inhibitory stimulation on interhemispheric balance. Additionally, this study will generate new data on how lesion localization and brain connectivity influence individual responses to rTMS, explaining variability in treatment efficacy.

Background and aims In the United States, 18 to 66% of individuals who experience a stroke will suffer from language impairment, referred to as aphasia. Aphasia profoundly impacts quality of life by restricting participation in social, familial, and professional spheres. While the gold standard for aphasia rehabilitation is Speech and Language Therapy (SLT), it often does not lead to complete recovery. Repetitive Transcranial Magnetic Stimulation (rTMS) or Transcranial Direct Current Stimulation (tDCS) are non-invasive brain stimulation techniques that could enhance post-stroke aphasia recovery. Most rTMS and tDCS studies in post-stroke aphasia following left hemispheric strokes have applied inhibitory stimulation targeting the right Inferior Frontal Gyrus (IFG) or facilitatory stimulation targeting the left IFG (i.e., Broca's area and its right homologue).

Aim 1: Focusing solely on the IFG for all participants with post-stroke aphasia may be inefficient, as the language system is widespread and complex, and involves large, interconnected networks also linked to non-linguistic cognitive functions such as working memory or executive functions. Previous research indicated that individuals with different language impairments may have varied recovery paths, and using the same cortical target for non-invasive brain stimulation might be ineffective or even impair recovery. Interestingly, rTMS targeting the motor cortex of the lips in people with phonological impairments improved accuracy and decreased phonological errors in a Picture Naming Task. This improvement likely resulted from rTMS modulating not only the targeted motor cortex but also the connected dorsal phonological stream via intra- and interhemispheric connections. Building on this principle, a pilot study was conducted using inhibitory rTMS targeting the right Anterior Temporal Lobe (ATL) in people with post-stroke aphasia suffering from lexicosemantic impairments, with improvement in Picture Naming for all participants. Lexicosemantic difficulties manifest as anomia and/or semantic errors (e.g., saying tea instead of coffee) and can result from deficits in lexical access or within the semantic system itself. Lexicosemantic processes involve a complex brain network. According to the hub-and-spokes theory, the ATL serves as a central amodal semantic hub that integrates information from various modalities. Other theories emphasize its multimodal and potentially lexical sensibility. This study hypothesizes that rTMS targeting the ATL will modulate the ventral stream of language, resulting in improvements in people with lexicosemantic impairments. This RCT aims to combine ATL stimulation with Semantic Feature Analysis (SFA) rehabilitation, which strengthens lexicosemantic links. This approach was expected to enhance the effect of rTMS on lexicosemantic processes.

Aim 2 and 3.3: Furthermore, the facilitation of the left hemisphere and the inhibition of the right undamaged hemisphere are based on the theory of transcallosal interhemispheric imbalance, where brain lesion leads to hyperactivation of the undamaged hemisphere and hypoactivation of the damaged hemisphere. The undamaged right hemisphere could play a maladaptive role in post-stroke aphasia recovery. However, the role of right hemisphere activation in post-stroke aphasia recovery remains unclear, as it may support recovery in certain participants during both the subacute and chronic phases. More research is needed to better understand the underlying mechanisms of facilitatory and inhibitory stimulation targeting respectively the undamaged and damaged hemispheres.

Aim 3: While non-invasive brain stimulation has been shown to be effective in post-stroke aphasia, certain people with post-stroke aphasia respond better than others. The previous sections highlighted the impact cortical target and stimulation parameters, but recent studies have also proposed that participant's aphasia type or lesion localization may influence their response. Further research, particularly using rTMS, is needed to clarify how lesion localization and disconnections contribute to recovery, as these factors may explain chronic post-stroke language impairments and predict rTMS efficacy. Since the ATL serves as a central hub in the ventral lexicosemantic stream, and rTMS modulates other brain areas through both inter- and intra-hemispheric connectivity, the hypothesis was as follow: a preserved connectivity between the ATL and the posterior temporal lobe will be a strong predictor of rTMS effectiveness.

Study Intervention Participants will be randomly assigned to one of three groups: (1) the inhibitory continuous Theta Burst Stimulation (cTBS) targeting the right ATL, (2) the facilitatory intermittent Theta Burst Stimulation (iTBS) targeting the left ATL, or (3) the sham control TMS group, with sham stimulation targeting either the left or right ATL in half of the participants. The intervention will consist of two weeks of stimulation according to group assignment, five times per week (10 sessions total), followed by 30 minutes of Speech-Language Therapy using the Semantic Feature Analysis method.

Inclusion Criteria:

• Right-handedness (assesed using the Edinburgh Handedness Inventory)
• Native English speakers
• Single left hemispheric stroke
• Diagnosis of aphasia and lexicosemantic impairments (e.g., anomia, semantic errors in picture naming or spontaneous speech) confirmed by a Speech and Language Pathologist or Physician

Exclusion Criteria:

• Severely impaired auditory comprehension (WAB Comprehension score of 0 or 1)
• Major phonological impairments (defined as > 2 errors in non-word repetition of one- and two-syllable items on the PALPA, 2)
• Contraindications to rTMS (e.g., uncontrolled epilepsy, history of seizures, intracranial ferromagnetic body, cochlear implant, or factors that lower seizure threshold (see Rossi et al., 2021)
• Contraindication to MRI
• Stroke involving the left ATL
• Severe depression or psychiatric history
• Illiteracy
• Known neurodegenerative disorders
• Severe uncorrected visual or hearing impairments

The expected data types include:

• Behavioral data: Accuracy and reaction times measures for cognitive and language tasks.
• Neuroimaging data: Structural MRI (T1 and T2 weighted) and functional MRI (resting state and task-based, for around 15 minutes per session)

Data will be made available upon publication of the respective study. Data will remain accessible for at least 10 years post-publication.

Interested research will be able to access the following data : - De-identified behavioral data (reaction time, accuracy, transcription) at be shared publicly using OSF

• Audio recordings will not be shared due to confidentiality risks
• Neuroimaging data (structural and functional MRI) will be shared after de-identification (face removal) in OpenNeuro
• Aggregate results will be disseminated via peer-reviewed publications and conference presentations.
• Study protocol, task descriptions, and analysis pipelines will be documented in manuscripts
• Metadata describing participant demographics (e.g., age, gender, stroke type, lesion location) will be included in shared dataset and available in manuscripts
• Documentation will be provided upon request to facilitate data interpretation