2023 Pilot Projects
Automatic Assessment of Neuropsychiatric Symptoms
Principal Investigator:
Ehsan Adeli, PhD
Stanford University
Aim 1: Adopt Ambient Intelligence (AmI) Algorithms to Correlate with positive affect. We will use our CV algorithms to extract facial expressions and behavior features. The data-driven features identified by our CV algorithm will be significantly correlated with positive affect SAM scores. This algorithm will also identify individual’s norms, i.e., frequencies of affects throughout the day in different places of the home. Aim 2: Examine the relationship between positive affect-related features and mood vs. agitation subsymdromes in NPS. We will correlate mood or agitation subdomains from MBI-C and NPI with our positive affect features. We expect positive affect features to be more related to mood than agitation, and to contribute additionally to differentiate mood vs. agitation from CV features collected in the parent study. The ultimate goal of this project is to generate data and infrastructure for a large-scale study with a possible clinical trial within the context of an R01. In parallel, to translate the technology into clinical practice, we will work with the Stanford Office of Technology Licensing (OTL) to productize the outcomes.
Collaborators:
Christine E. Gould, PhD, Stanford University
Silvia Tee, MD,
Stanford University
Developing a novel model to study the effects of aging and Alzheimer’s disease pathology on complex sociocognitive traits
Principal Investigator:
Aim 1: Delineate the trajectory of bonding and response to loss across aging in prairie voles. We will examine bonding and loss at 6, 12, and 18 months of age. Based on previous literature in humans and voles indicating maintenance or improvement of social bonds across aging, we predict that age will increase bond strength, facilitate the emergence of reliable behavioral patterns in paired voles, and result in prolonged bond persistence after partner separation as a way to operationalize loss. Furthermore, by assessing behavior of both individuals within a dyad and applying novel pose estimation (DeepLabCut21) and behavioral classification (SimBA22) approaches, we expect to uncover unique intra-pair behavioral dynamics that emerge over the course of aging. Aim 2: Establish methods for the introduction of AD pathology in prairie voles. We will establish methods to introduce brain-wide and region-specific AD-relevant pathology in prairie voles, which we will map over the course of acute aging (up to 6 months). This will establish the first prairie vole model of AD to be leveraged in future studies of pathology-induced circuit dysfunction underlying social and emotional deficits at individual and dyadic levels. Together these aims will establish prairie voles as a novel model for interrogating the effects of normal and abnormal aging on complex socioemotional phenotypes.