Abstract:
Delirium is a frequent and serious complication in the acute phase after stroke, yet reliable predictors of its occurrence are lacking. Emerging evidence implicates disturbed large-scale brain network dynamics as a key mechanism leading to delirium. Transcranial magnetic stimulation combined with electroencephalography (TMS-EEG) provides a unique approach to assess cortical network responsiveness and response complexity in vivo.
In this study we investigated to what extent the TMS-EEG features, especially fast perturbational complexity index (PCIST), a TMS-EEG measure of response integration and differentiation, predict the development of PSD when tested within the first 48 hours after stroke onset but prior to delirium onset.
We prospectively enrolled 34 acute stroke patients admitted to the stroke unit. In addition to resting-state EEG, TMS-EEG were recorded by targeting the dorsolateral prefrontal cortex (DLPFC) and superior parietal lobule (SPL) of the ipsilesional and contralesional hemisphere. Delirium was assessed every 8 hours using the Intensive Care Delirium Screening Checklist (ICDSC) and DSM-5 criteria. PCIST and natural frequency were extracted from TMS-evoked EEG responses; resting-state EEG spectral power was also analyzed. Predictive performance was evaluated using logistic regression, ROC analysis, and correlations with clinical and imaging markers, including NIHSS and lesion volume.
In our study, thirteen patients (38%) developed delirium. Patients in PSD group had higher NIHSS score, larger lesion volume, and increased theta-band power in resting-state EEG, compared with the non-PSD group. However, these measures showed limited predictive ability for PSD. In TMS-EEG–based analyses, although we observed a reduction in the natural frequency of the ipsilesional DLPFC in the PSD group, its predictive value for PSD was even lower than that of conventional clinical indicators.
Most importantly, this study identifies PCIST, a perturbation-based marker of cortical complexity, as a robust and independent predictor of PSD. PCIST reduction was observed across stimulation sites and hemispheres, reflecting a network-level vulnerability that transcends focal lesion characteristics. These findings support the hypothesis that PSD arises from a collapse in thalamo-cortical integrative dynamics—disrupting the brain’s capacity for reentrant processing and coherent network communication.
PCIST proved feasible and reproducible under real-world stroke unit and ICU conditions, confirming its clinical applicability for early delirium risk stratification. Beyond prediction, PCIST may also serve as a mechanistically grounded target for individualized neuromodulatory interventions. Future work should focus on simplifying the TMS–EEG pipeline for routine bedside use and testing whether PCIST-guided stimulation protocols can prevent cognitive deterioration in high-risk stroke patients.
Together, these results advance PCIST from a theoretical construct to a translational tool—offering a novel avenue for precision medicine approaches to delirium prevention in acute neurocritical care.
Post-stroke delirium