Abstract Background: Unilateral spatial neglect (USN) is a frequent post-stroke disorder of spatial attention that impairs daily function and predicts poorer recovery. Non-invasive neuromodulation like repetitive transcranial magnetic stimulation (rTMS) can improve USN, but the most effective stimulation strategy is still unknown. Objective: This randomized study compared low-frequency rTMS to the contralesional hemisphere versus high-frequency rTMS to the ipsilesional hemisphere in stroke patients with USN, based on interhemispheric inhibition theory, to determine which protocol yields superior recovery. Methods: Fifty-three patients with right-hemisphere stroke and USN were randomized into two groups: (1) low-frequency rTMS (1 Hz) applied to the left posterior parietal cortex (P3) n=27, and (2) high-frequency rTMS (10 Hz) applied to the right posterior parietal cortex (P4) n=26. All patients received standard rehabilitation for four weeks. Outcome measures before and after treatment included spatial neglect (Behavioral Inattention Test, BIT, and Catherine Bergego Scale, CBS), motor function (Fugl-Meyer Assessment, FMA), and activities of daily living (Modified Barthel Index, MBI). We compared pre- and post-treatment changes within each group and between groups using appropriate statistical tests. Results: After 4 weeks, both groups improved significantly on all spatial neglect measures (BIT subtests and total score increased; CBS score decreased; all pConclusion: Both contralesional low-frequency and ipsilesional high-frequency rTMS substantially improved neglect, motor recovery, and ADL performance. Consistent with the interhemispheric-inhibition model, high-frequency stimulation of the lesioned hemisphere produced larger neglect improvements and ADL gains than low-frequency stimulation of the intact hemisphere. These findings suggest that ipsilesional excitatory rTMS may be more effective for early neglect rehabilitation than contralesional inhibitory rTMS. Trial registration: ChiCTR2500106020, retrospective registration date July 16, 2025.
Peng et al. (Tue,) studied this question.