Abstract T P122: Dose-Response Relationships in Brain and Behavioral Measures with Stroke Rehabilitation Using a Brain-Computer Interface

Abstract

Introduction: Brain-computer interface (BCI) is an emerging technology for stroke rehabilitation, but little is known about how the administration of these therapies affects brain and behavior. We examine relationships between BCI therapy dose, intensity, and frequency with changes in laterality index (LI), Action Research Arm Test (ARAT), and Stroke Impact Scale (SIS). Methods: We collected ARAT and SIS scores as well as anatomical and functional MRI of 16 stroke patients with upper extremity motor impairment before, during, and after therapy using a BCI system with tongue and functional electrical stimulation. We acquired fMRI during finger tapping of each hand and computed LI with 3 mask sets at 2 thresholds. Changes from baseline LI and behavioral scores were assessed for correlation with cumulative therapy sessions, cumulative BCI runs, and overall runs/session using Spearman analysis and generalized estimating equations (GEE). Changes from baseline were also compared between subjects with different therapy frequencies. Results: Increased sessions correlated with changes in impaired finger tapping LI in 2 of 6 mask-threshold combinations (p<0.05) and trended to significance (0.05≤p<0.1) in 2 additional mask-threshold combinations. The relationship between cumulative runs and changes in impaired finger tapping LI also trended to significance (0.05≤p<0.1) in 2 mask-threshold combinations. Changes in SIS Strength correlated (p<0.05) with cumulative runs and with runs/session. All 8 of these relationships identified with Spearman analysis were significant with analysis using GEE (p < 0.05). No relationships were observed between therapy dose or intensity and ARAT scores or unimpaired hand measures. No differences were observed between subjects with higher and lower therapy frequency. Conclusions: Patterns of brain activation shift toward contralesional hemisphere with additional therapy sessions and additional BCI runs. Increased BCI therapy intensity also correlates with increased self-reported strength. When using BCI in stroke rehabilitation, therapy frequency may be less important than dose and intensity, and neuroimaging and self-report measures may be more sensitive to differences in therapy administration.