RuCCS - Dr. Albert Powers - In Person Talk - Tuesday, April 07, 2026

Value of the Information & Strength of the Argument

The talk provides valuable insights into the computational underpinnings of hallucinations, presenting a clear and testable hypothesis. The argumentation is solid, building from a well-established framework (Bayesian perception) to a specific prediction about hallucinations, and then testing it with a carefully designed experiment. The speaker acknowledges alternative explanations (e.g., response bias) and addresses them with confidence ratings and neural data. The use of a 2x2 design to dissociate hallucinations from psychosis is methodologically strong. The talk also places the work in historical context, citing early conditioning experiments, which adds depth. However, the talk is primarily a presentation of the speaker’s own research, and while it references other work, it does not provide a comprehensive review of the field.

Scientific Rigor, Source Quality, Title Accuracy

The speaker demonstrates scientific rigor by grounding his work in established theories (e.g., Helmholtz’s unconscious inference) and by using rigorous psychophysical methods (e.g., QUEST procedure). He cites specific studies, such as Seashore (1895) and Ellson (1930s), and mentions contemporary researchers like Ralph Hoffman and Michel Hamson. The title accurately reflects the content, as it is a recorded talk by Dr. Powers at RuCCS. The talk is an expert opinion, but it is based on peer-reviewed research and presents original data. The speaker also discusses limitations and alternative interpretations, which enhances credibility. No comments were provided for analysis.

The title accurately reflects the content: a recorded in-person talk by Dr. Albert Powers at RuCCS, focusing on computational psychiatry and psychosis.

8/10

The talk is given by a psychiatrist and neuroscientist at Yale, with a joint appointment in psychiatry and psychology, and is based on peer-reviewed research. The speaker presents a clear computational framework and references historical and contemporary studies. However, as a single expert talk, it lacks the breadth of a systematic review or meta-analysis.

Markers derived by PSI from the transcript: the creator did not define chapters.

• 02:00Introduction of Dr. Albert Powers by the host, highlighting his role at Yale and his research on psychosis.
• 05:00Dr. Powers outlines the talk: medical nosology, psychiatric nosology, auditory hallucinations as an example, and broader implications.
• 10:00Discussion of medical nosology using jaundice as an example, emphasizing the importance of understanding causal mechanisms.
• 15:00Introduction of the Bayesian framework for perception, explaining priors, likelihoods, and precision.
• 20:00Hypothesis that hallucinations arise from overweighted priors, and historical experiments by Seashore and Ellson.
• 30:00Description of the 2x2 design (psychosis vs. no psychosis, hallucinations vs. no hallucinations) and the experimental paradigm.
• 40:00Presentation of behavioral results: individuals with hallucinations show more false alarms and higher confidence in false perceptions.
• 50:00fMRI results showing auditory cortex activity correlates with tone intensity, supporting the percept-like nature of hallucinations.
• 1:00:00Discussion of implications for understanding psychosis development and potential for prevention and treatment.
• 1:10:00Q&A session and concluding remarks.

• Seashore, C. E. (1895). Measurements of illusions and hallucinations in normal life. — Historical experiment on conditioning auditory hallucinations.
• Ellson, D. G. (1941). Hallucinations produced by sensory conditioning. — Follow-up experiments on conditioned hallucinations.
• Powers, A. R., et al. (2017). Pavlovian conditioning-induced hallucinations result from overweighting of perceptual priors. — The speaker's own study on hallucinations and Bayesian priors.

• Corlett, P. R., et al. (2019). Hallucinations and strong priors. — Supports the idea that hallucinations are related to altered prior weighting.

• Some studies suggest hallucinations may be due to impaired sensory evidence rather than overweighted priors. — Alternative computational accounts exist, but the speaker's data support the prior overweighting hypothesis.