Robotic Assessment of Neurocognitive Deficits in Adults with Temporal Lobe Epilepsy and Genetic Generalized Epilepsy

Abstract

Cognitive impairment from epilepsy is well-recognized. Neuropsychological assessment can detect cognitive impairments through clinical criteria and quantitative measures. Although expectations exist for cognitive variation derived from the diagnosis of epilepsy, phenotypic variation of cognitive deficits exists within epilepsy subtypes. Classification via severity and type of cognitive deficits have been shown across multiple studies in temporal lobe epilepsy (TLE). In this study, we test the validity and utility of Kinarm robotic assessment to measure cognitive ability beyond motor function in people with epilepsy by comparing it to a neurocognitive screening. Participants with temporal lobe epilepsy (TLE, n=33) and genetic generalized epilepsy (GGE, n=25) underwent a neurocognitive screening and Kinarm assessment. Pearson correlations were used to assess correlations between the two types of assessments’ similar neurocognitive domains, determined a priori. Correlations exist in the neurocognitive domains of complex attention (3/6 tests, p<0.05), executive function (3/7 tests, p<0.05), memory (2/4 tests, p<0.05), visual-motor coordination (5/12 tests, p<0.05), and visuospatial skill (1/7 tests, p<0.05) were moderate (r ~.30) to strong (r ~.50) between our brief neurocognitive assessments and robotic assessments. We further investigated the correlation between quality of life and Kinarm assessment, and seizure duration and Kinarm assessment and found no significant results in either set of tests. To test whether Kinarm assessment could replicate past research concerning epilepsy cognitive phenotyping, we applied a fuzzy c-means clustering method to our data. We found 3 clusters exist among people with TLE: minimal deficits (45% of participants with TLE), partial deficits (27%), and deficits across tasks (27%). We found 3 clusters exist among people with GGE: memory and executive deficits (40% of participants with GGE), processing speed deficits (36%), and complex motor deficits (24%). These clusters show similar patterns as previous research, which suggests 2 that robotic assessment may be able to assess cognitive phenotypes of epilepsy. Overall, these results demonstrate that robotic assessment can measure cognitive ability beyond sensorimotor function, much like a neurocognitive screening of people with generalized and focal epilepsies.