Evaluating a Novel Brain-Computer Interface and EEG Biomarkers For Cognitive Assessment in Children With Cerebral Palsy

Abstract

Standardized neuropsychological assessments and research instruments are typically administered with verbal queries, pictures and manipulatives that require verbal or motor responses. Thus, they are often inaccessible to people with physical and communicative impairments. The goal of this dissertation was to investigate alternative approaches that do not require any motor or speech input to assess cognitive capacity of an individual. The first approach involved using a brain-computer interface (BCI) that was adapted to facilitate the administration of a Peabody Picture Vocabulary Test (PPVT-IV). Which is a receptive vocabulary assessment than can be used as a proxy for intelligence. The second approach was to use brain dynamics such as functional connectivity and bandpass analysis to assess cognitive capacity of an individual. We then tested these two approaches on typically developing (TD) individuals (N=11) people with cerebral palsy (CP) (N=18). Our results suggest that children with cerebral palsy show signs of lower intelligence than typically developing children when using functional connectivity and power band analysis, however, they performed equally well in the PPVT-IV. We believe this is due to the neural compensation resulting from the subjects’ pathology. Thus, the preferred method for assessing cognitive measures in an individual with severe motoric impairments is a BCI. By using a BCI, a user can respond to standardized cognitive assessments that already have well established norms. However, it is important to make sure that when designing these systems, the changes made to adapt the cognitive assessment for the BCI do not alter the format and psychometrics of the test. Our BCI able to maintain the psychometrics of a PPVT-IV test and perform with an accuracy of 97.78 ± 4.06. In addition, scores on the BCI-facilitated PPVT-IV and the standard PPVT-IV were highly correlated (r = 0.95, p<0.001) with a mean difference of 2.0 ± 6.4 points, which is within the standard error of the PPVT-IV. Thus, our BCI-facilitated PPVT-IV provided comparable results to the standard PPVT-IV, suggesting that populations for whom standardized cognitive tests are not accessible could benefit from our BCI-facilitated approach.