Event Details:
Uncovering mental structure through data-driven ontology discovery
Ian Eisenberg
Mind, Brain, Computation and Technology graduate trainee, Stanford University
Abstract
Cognitive neuroscience has linked neural activity to a wealth of cognitive processes, yet struggles to produce a cumulative account of neural function. This slow progress has many causes, but is partially explained by the lack of systematic ontologies describing brain structure and mental function. While integrative brain atlases have been steadily improving, commensurate efforts to improve cognitive ontologies have been limited. We address this by developing a data-driven cognitive ontology derived from individual differences across a broad range of behavioral tasks, self-report surveys, and real-world outcomes. 522 participants completed 62 different measures on Mechanical Turk related to decision-making, working memory, cognitive control, impulsivity, and personality, amongst other psychological constructs. Interestingly, though subsets of the tasks and surveys putatively reflect similar constructs, we find that they bifurcate in the ontology. Using exploratory factor analysis, we identify two low-dimensional cognitive spaces that separately capture behavioral tasks and surveys. Hierarchical clustering within these spaces identify sensible clusters which capture psychological "kinds", related to, but separate, from the dimensions identified with factor analysis. Overall, this structure discovery reveals a simpler cognitive ontology than typically employed in the psychological sciences. As real-world relevance is an essential feature of theoretical constructs, we also evaluated whether tasks and surveys can predict real-world outcomes. We reduced the self-reported real-world outcomes to 8 "target" factors (e.g. mental health, binge drinking), computed individual factor scores, and assessed predictive ability using cross-validated ridge regression. While surveys performed moderately well, tasks showed almost no predictive ability.
Cognitive ontologies describe the psychological constructs through which most human neuroscience is understood. We demonstrate that data-driven structure discovery techniques can profitably improve these ontologies.
Related Video:
Uncovering mental structure through data-driven ontology discovery
Interactions between goal states, attention, and memory in episodic remembering
Kevin Madore, PhD
Postdoctoral Research Fellow in Psychology, Stanford University
Abstract
Moment-to-moment interactions between goal states, attention, and episodic memory retrieval may influence when individuals remember and when they forget. We recorded concurrent scalp EEG and pupillometry during an encoding/retrieval paradigm with 80 young adults to examine how (a) multimodal indices of attentional lapses (e.g., pre-trial increases in alpha power and decreases in theta power and pupil diameter) relate to goal-state representation and episodic memory, and how (b) trait differences in sustained attention and related constructs may contribute to these relationships. At encoding, participants performed two tasks, classifying individual objects on either a conceptual or perceptual dimension. At retrieval, participants oriented to one of three retrieval goals (concept before? vs. percept before? vs. new item?) and then were cued with an old or new object and made the retrieval judgment. We examined how trial-by-trial tonic fluctuations in alpha (8-12Hz) and theta (4-7Hz) oscillatory power and in pupil diameter (a) before orienting to the retrieval goal and (b) before viewing the object and making the retrieval judgment affect accuracy. Across goal states, retrieval performance was predicted by fluctuations in alpha and theta power and pupil diameter during the 1000ms before goal orienting and during the 1000ms before the retrieval cue. Pre-goal and pre-retrieval mean alpha power were significantly higher for misses relative to hits. In addition, mean pre-goal theta power and pupil diameter were significantly lower for misses relative to hits. Pre-retrieval variability in pupil diameter also predicted false alarms relative to correct rejections. Individual difference analyses further revealed that (a) self-reported levels of media multitasking, self-reported rates of spontaneous mind wandering, and behaviorally assayed commission error rates on the gradual-onset continuous performance test (gradCPT) significantly positively related to subject-level EEG and pupil metrics during the pre-goal and pre-retrieval epochs of the separate memory task. Media multitasking, commission error rates on the gradCPT, and subject-level EEG and pupil metrics were also significantly negatively related to d’ (overall memory performance). These results highlight how preparatory attention and goal-state representation impact episodic remembering at the state and trait levels.
Related Video:
Interactions between goal states, attention, and memory in episodic remembering