Validation of the Dynamic Affect Recognition Test (DART): Behavior and Anatomy in Neurodegenerative Disease (P6.237)

OBJECTIVE: Evaluate a novel video-based emotion measure in neurodegenerative patients; demonstrate that deficits predict atrophy in emotion-relevant brain systems. BACKGROUND: Behavioral variant frontotemporal dementia (bvFTD) patients show disproportionately impaired emotion comprehension, yet traditional picture-based emotion tests show deficits in patients without real-life impairment. The Dynamic Affect Recognition Test (DART) is a novel tablet-based test designed to quickly assess emotion comprehension with ecologically valid stimuli. DESIGN/METHODS: 146 participants (20 bvFTD, 51 Alzheimer’s disease (AD), 19 semantic variant primary progressive aphasia (svPPA), 8 progressive non-fluent aphasia (nfvPPA), 9 progressive supranuclear palsy (PSP), 39 healthy controls (NC)), watched 12 20-second videos of an actor expressing emotions (happy, surprised, sad, angry, fearful, disgusted) via congruent facial/vocal/postural cues, with semantically neutral scripts. Subjects identified the emotion from a randomized visual array. Videos were FACS-coded to ensure valid emotional expression. Dartel-based (SPM8) VBM of T1 MRI was performed in a covariates-only analysis of DART score controlling for sex, age, MMSE, and TIV. RESULTS: Group differences in sex, age, and MMSE warranted inclusion of these factors as confounds. SAS proc GLM modeling showed diagnostic group explained 47% of the variance in DART score, with svPPAs (6.07±0.48,p<0.0001), bvFTDs (6.62±0.50,p<0.0009), and PSPs (7.27±0.68,p<0.0378) performing significantly worse than controls, while ADs and PNFAs performed normally. As predicted, VBM showed poorer DART performance correlated with focal volume loss only in emotion-related areas including bilateral medial temporal, temporal pole, anterior insula, basal ganglia, and ventromedial frontal regions (pFWE<0.05). CONCLUSIONS: We have developed and validated a quick, tablet-based, video test of emotion reading that matches the expected structural anatomy of emotion, on which diagnostic groups known for socioemotional deficits fail, while those with intact real-life emotion reading perform normally despite cognitive deficits. Tests allowing quick but valid evaluation of emotion reading deficits at the bedside are essential for full neuropsychological evaluation of neurodegenerative disease. Disclosure: Dr. Radke has nothing to disclose. Dr. Adhimoolam has nothing to disclose. Dr. Shany-Ur has nothing to disclose. Dr. Possin has nothing to disclose. Dr. Marin has nothing to disclose. Dr. Zackey has nothing to disclose. Dr. Poorzand has nothing to disclose. Dr. Miller has received personal compensation in an editorial capacity for Cambridge University Press, Guilford Publications, Inc., and Neurocase. Dr. Rankin has nothing to disclose.