About

Dr. Albert Powers is an Associate Professor at the Yale University Department of Psychiatry and serves as the Medical Director and Associate Director of the PRIME Psychosis Risk Clinic at Yale. His clinical work involves treating individuals experiencing symptoms of early psychosis. Dr. Powers employs computational approaches to investigate how sensory systems may malfunction, leading to hallucinations and other psychosis symptoms. His research integrates clinical practice with computational methods to better understand the underlying mechanisms of psychotic disorders, aiming to improve diagnosis and treatment strategies for early psychosis.

Research topics

• Psychology
• Cognitive psychology
• Psychiatry
• Medicine
• Clinical psychology

Selected publications

• Threading the Needle: Practical Considerations for Merging Theory-Driven Computational Psychiatry With Data-Driven Analytics to Enhance Precision Health at Scale

  Biological Psychiatry Cognitive Neuroscience and Neuroimaging · 2026-02-01

  articleOpen access

  The rapidly evolving field of computational psychiatry enables quantification of specific cognitive processes and their underlying mechanisms in a translational and potentially scalable manner, using a combination of data collection via mechanistically informed behavioral tasks and theory-driven mathematical modeling. In parallel, transdiagnostic, dimensional approaches to psychiatric diagnostics, such as the Research Domain Criteria and Hierarchical Taxonomy of Psychopathology, seek to facilitate links between clinical research and real-world clinical reality, which rarely respects traditional diagnostic boundaries. These two approaches are seldom combined. In addition, while most psychiatric disorders are defined by their longitudinal course, our ability to predict symptom trajectories and tailor treatments to the individual remains limited, in part due to a dearth of longitudinal data collected using assessments sensitive to individual change over time. To address these gaps, the recently launched IMPACT-Y (Individually Measured Phenotypes to Advance Computational Translation at Yale) study is collecting longitudinal data from a transdiagnostic cohort of 2400 individuals, using a combination of traditional clinical research methods (e.g., health records, standardized assessments) and more novel computational approaches (e.g., behavioral tasks with demonstrated sensitivity to latent constructs and to within-person change, spoken narrative data). Here, we discuss unique challenges and opportunities in study design and analysis considerations of IMPACT-Y. Incorporating both theory- and data-driven analytics, we hope that IMPACT-Y will provide an unprecedented resource for characterizing longitudinal trajectories of core computational psychiatry constructs (e.g., reward learning) within and between individuals for parsing heterogeneity beyond traditional diagnostic categories and for linking inter- and intraindividual clinical variability to underlying mechanisms.

  PublisherDOI

• Computational phenotypes underlying effort-based decision-making and negative symptoms in a transdiagnostic severe mental illness sample

  Molecular Psychiatry · 2026-02-14

  article
  PublisherDOI

• 20. Upstream Sensory Disruptions Relate to Hallucination Propensity in Dementia With Lewy Bodies

  Biological Psychiatry · 2026-04-25

  articleSenior author
  PublisherDOI

• The Perception of Visual Distortions in Virtual Reality and Their Association with Psychosis-Relevant Traits

  2026-01-24

  articleOpen access

  Background Visual distortions (VDs) are frequently reported but insufficiently characterized in early psychosis and may reflect alterations in perceptual processing. Computational accounts emphasize aberrant precision-weighting of sensory evidence and prior expectations, but most empirical work has used simplified laboratory stimuli. It remains unclear whether these mechanistic accounts generalize to more naturalistic perceptual contexts. MethodsWe developed virtual reality (VR) tasks simulating VDs in photorealistic scenes and explored associations between VD perception and psychosis-relevant traits in a non-clinical sample. 120 healthy adults completed three tasks designed to probe complementary aspects of perceptual inference: (1) a naïve detection task (NDT) assessing expectation-driven false percepts of unspecified visual changes; (2) a perceptual uncertainty task (PUT) testing how detection under ambiguous sensory evidence is shaped by recent trial history; and (3) an adaptive two-interval forced-choice task (2-IFC) quantifying bottom-up sensitivity to psychosis-like VDs. Psychosis-relevant traits were assessed using validated questionnaires of delusional ideation, anomalous perceptual experiences, and schizotypal traits. ResultsHigher levels of delusional ideation and schizotypy were associated with stronger expectation effects, reflected in increased false percepts in the NDT trial without changes and enhanced detection of near-threshold distortions when these trials followed distortion-present trials in the PUT. These traits were also linked to reduced 2-IFC perceptual sensitivity to VDs, particularly brightness changes.ConclusionsThese findings demonstrate the feasibility of VR-based VD assessment and provide initial evidence that individual differences in VD perception relate to psychosis-relevant traits. Results suggest contributions from both diminished sensory sensitivity and increased reliance on expectations in naturalistic perceptual contexts.

  PublisherDOI

• The Perception of Visual Distortions in Virtual Reality and Their Association with Psychosis-Relevant Traits

  2026-01-28

  articleOpen access

  Background Visual distortions (VDs) are frequently reported but insufficiently characterized in early psychosis and may reflect alterations in perceptual processing. Computational accounts emphasize aberrant precision-weighting of sensory evidence and prior expectations, but most empirical work has used simplified laboratory stimuli. It remains unclear whether these mechanistic accounts generalize to more naturalistic perceptual contexts. MethodsWe developed virtual reality (VR) tasks simulating VDs in photorealistic scenes and explored associations between VD perception and psychosis-relevant traits in a non-clinical sample. 120 healthy adults completed three tasks designed to probe complementary aspects of perceptual inference: (1) a naïve detection task (NDT) assessing expectation-driven false percepts of unspecified visual changes; (2) a perceptual uncertainty task (PUT) testing how detection under ambiguous sensory evidence is shaped by recent trial history; and (3) an adaptive two-interval forced-choice task (2-IFC) quantifying bottom-up sensitivity to psychosis-like VDs. Psychosis-relevant traits were assessed using validated questionnaires of delusional ideation, anomalous perceptual experiences, and schizotypal traits. ResultsHigher levels of delusional ideation and schizotypy were associated with stronger expectation effects, reflected in increased false percepts in the NDT trial without changes and enhanced detection of near-threshold distortions when these trials followed distortion-present trials in the PUT. These traits were also linked to reduced 2-IFC perceptual sensitivity to VDs, particularly brightness changes.ConclusionsThese findings demonstrate the feasibility of VR-based VD assessment and provide initial evidence that individual differences in VD perception relate to psychosis-relevant traits. Results suggest contributions from both diminished sensory sensitivity and increased reliance on expectations in naturalistic perceptual contexts.

  PublisherDOI

• Emergence and dynamics of delusions and hallucinations across stages in early psychosis

  UNC Libraries · 2026-02-02

  articleOpen access
  PublisherDOI

• Compensatory hallucinogenesis across three neuropsychiatric disorders: a Bayesian account

  Brain Communications · 2026-01-01

  articleOpen accessSenior author

  Emerging evidence suggests that hallucinations may arise because of an over-reliance on prior knowledge during perception. While best established in psychosis-spectrum illness, data also support the presence of this abnormality in other hallucination-prone neuropsychiatric illnesses that vary in their association with disruption of sensory circuits. In this piece, we ask whether an over-weighting of expectations may be conceived of as a compensatory response to degraded incoming sensory information. We make the case that visual hallucinogenesis across a wide array of neuropsychiatric disorders can be captured within a common Bayesian computational framework, as a compensatory response to sensory signal disruptions at different levels of the visual processing hierarchy. We focus on three specific disorders (Charles Bonnet syndrome, dementia with Lewy Bodies and psychosis) with prominent visual hallucinations and highlight the fact that these disorders describe a spectrum of visual impairment where the overtness and localization of the visual processing disruption is reflected in the characteristics of the emergent visual hallucinations. We examine how discrete sensory disruptions in Charles Bonnet syndrome translate to hallucinations via known circuits, and then how different disruptions in dementia with Lewy Bodies and Schizophrenia may lead to hallucinations with distinct phenomenology, comorbidities and circuit involvement. Finally, we appeal to emerging computational theories to unite these observations under a common conceptual umbrella. Taken together, this work presents a means of understanding how sensory disruptions could interact with other aspects of cognitive and neural architecture to produce hallucinations across neuropsychiatric disease. It is our hope that this framework will help in efforts to identify pathophysiologically distinct patient subgroups and new pharmacological and circuit-based interventions.

  PublisherDOI

• Cognitive assessment in the Accelerating Medicines Partnership® Schizophrenia Program: harmonization priorities and strategies in a diverse international sample

  UNC Libraries · 2025-12-05

  articleOpen access
  PublisherDOI

• Sampling from different populations: Sociodemographic, clinical, and functional differences between samples of first episode psychosis individuals and clinical high-risk individuals who progressed to psychosis

  UNC Libraries · 2025-12-05

  articleOpen access
  PublisherDOI

• The electroencephalography protocol for the Accelerating Medicines Partnership® Schizophrenia Program: Reliability and stability of measures

  UNC Libraries · 2025-12-05

  articleOpen access

  Individuals at clinical high risk for psychosis (CHR) have variable clinical outcomes and low conversion rates, limiting development of novel and personalized treatments. Moreover, given risks of antipsychotic drugs, safer effective medications for CHR individuals are needed. The Accelerating Medicines Partnership® Schizophrenia (AMP® SCZ) Program was launched to address this need. Based on past CHR and schizophrenia studies, AMP SCZ assessed electroencephalography (EEG)-based event-related potential (ERP), event-related oscillation (ERO), and resting EEG power spectral density (PSD) measures, including mismatch negativity (MMN), auditory and visual P300 to target (P3b) and novel (P3a) stimuli, 40-Hz auditory steady state response, and resting EEG PSD for traditional frequency bands (eyes open/closed). Here, in an interim analysis of AMP SCZ EEG measures, we assess test-retest reliability and stability over sessions (baseline, month-2 follow-up) in CHR (n = 654) and community control (CON; n = 87) participants. Reliability was calculated as Generalizability (G)-coefficients, and changes over session were assessed with paired t-tests. G-coefficients were generally good to excellent in both groups (CHR: mean = 0.72, range = 0.49–0.85; CON: mean = 0.71, range = 0.44–0.89). Measure magnitudes significantly (p < 0.001) decreased over session (MMN, auditory and visual target P3b, visual novel P3a, 40-Hz ASSR) and/or over runs within sessions (MMN, auditory/visual novel P3a and target P3b), consistent with habituation effects. Despite these small systematic habituation effects, test-retest reliabilities of the AMP SCZ EEG-based measures are sufficiently strong to support their use in CHR studies as potential predictors of clinical outcomes, markers of illness progression, and/or target engagement or secondary outcome measures in controlled clinical trials.

  PublisherDOI

Recent grants

• Toward a Computationally-Informed, Personalized Treatment for Hallucinations

  NIH · $404k · 2020–2024

• A Hearing Test for Hallucinations: Toward Development of Computational Markers for Early Diagnosis

  NIH · $983k · 2018–2023

• NIH Grant F30DC009759

  NIH · $112k · 2013

Frequent coauthors

• Scott W. Woods

  Yale University

  80 shared

• Catalina Mourgues

  Yale University

  48 shared

• Philip R. Corlett

  Yale University

  46 shared

• Victoria Fisher

  Yale University

  40 shared

• Armin Drusko

  University Hospital Heidelberg

  38 shared

• David Baumeister

  University Hospital Heidelberg

  38 shared

• Thomas Graven‐Nielsen

  Aalborg University

  37 shared

• Megan Elizabeth McPhee Christensen

  University Hospital Heidelberg

  36 shared

Labs

• Powers LabPI