Assessing the correlation between sensorimotor evoked potential and cognitive impairment in post-stroke patients

Background Stroke is a leading cause of death and disability worldwide, with 35–80% of survivors developing cognitive impairment that affects quality of life. Current diagnostic tools are limited, highlighting the need for objective, cost-effective biomarkers for post-stroke cognitive impairment (PSCI).Aim To investigate the relationship between PSCI severity and sensorimotor evoked potentials, compare interhemispheric sensorimotor cortex excitability, and assess the utility of these potentials for PSCI screening.Subjects and methods Twenty-two stroke patients underwent neuropsychological and clinical evaluations at 3–6 months post-stroke, with testing performed within 7±3 days of the scheduled date. Patients were divided into mild cognitive impairment (MCI) and dementia groups. Motor evoked potentials (MEP) elicited by single-pulse transcranial magnetic stimulation assessed primary motor cortex excitability, while somatosensory evoked potentials (SEP) from median nerve stimulation evaluated primary somatosensory cortex excitability.Results MEP absence showed a dose-dependent relationship with cognitive severity (Kendall’s tau-b = −0.467, p = 0.029); 80% of MEP-absent patients had dementia versus 20% of MEP-present cases. In hemispheric analysis, affected hemispheres displayed elevated the resting motor threshold (RMT) (p < 0.001) and prolonged SEP N20 latency (p < 0.05), while unaffected hemispheres showed increased RMT in the MCI-with-MEP and dementia-without-MEP groups (p < 0.05), with reduced amplitude in dementia-with-MEP patients (p < 0.01). No linear SEP-cognition relationship emerged.Conclusions PSCI severity correlates with MEP presence/absence, not SEP measures. Notably, the excitability of the unaffected motor cortex progressively declines with worsening cognition. Although MEP shows promise as a biomarker for cognitive impairment assessment, longitudinal validation is needed to confirm its predictive utility.