# Songmaking in a Group (SING): Music, Hallucinations & Predictive Coding

> **NIH NIH R33** · YALE UNIVERSITY · 2023 · $487,783

## Abstract

Abstract
People with psychotic illnesses perceive and believe things about themselves, the outside world
and other people that do not obtain. This can be very distressing for them, their family members
and friends. Listening to and performing music can help mitigate this distress, but we do not
know why. This project aims to find out. Perceiving and believing, about self and others, is
achieved by making predictions and updating those predictions in light of new evidence,
particularly if that new evidence is very reliable or precise. One way that music might help
psychosis involves this precision. By making one set of predictions more precise—predictions
about music— other predictions can change. This might be why we tap our toes or sing along to
music we enjoy. We propose that experiencing more reliable predictions about ones’ actions (by
singing) and other people (by singing along with them) will help to change the predictions that
underwrite the symptoms of psychosis. We will test whether this is true in an initial R61 study,
tracking the change in performance of a series of prediction-based tasks as a result of musical
experience by prosecuting three specific aims: Specific Aim 1 will examine the impact of song-
making in a group (SING) on conditioned hallucination task performance, a procedure that
safely and reliably engenders hallucinations in the laboratory. We predict SING will reduce the
number and mechanisms of hallucinations in the laboratory. Specific Aim 2 will examine how
social learning changes with SING. Using a reputation learning task, we will measure social
learning rates. We predict they will increase with SING. Specific Aim 3 will examine
participants’ subjective experience of self-hood and how they change with SING using
computational linguistic analysis. We predict SING will decrease linguistic markers of distress. If
those studies prove successful, we will – in a follow-up R33 study – use metrics from the R61 to
decompose the musical intervention into its key ingredients – asking whether it is important to
be active (or merely passively experience music), and whether owning the music produced is
important to its impact on precision of processing. These studies will help us refine whether,
how and to whom we deliver musical intervention for serious mental illness.

## Key facts

- **NIH application ID:** 10704492
- **Project number:** 5R33MH123028-04
- **Recipient organization:** YALE UNIVERSITY
- **Principal Investigator:** PHILIP Robert CORLETT
- **Activity code:** R33 (R01, R21, SBIR, etc.)
- **Funding institute:** NIH
- **Fiscal year:** 2023
- **Award amount:** $487,783
- **Award type:** 5
- **Project period:** 2019-09-10 → 2025-07-31

## Primary source

NIH RePORTER: https://reporter.nih.gov/project-details/10704492

## Citation

> US National Institutes of Health, RePORTER application 10704492, Songmaking in a Group (SING): Music, Hallucinations & Predictive Coding (5R33MH123028-04). Retrieved via AI Analytics 2026-05-24 from https://api.ai-analytics.org/grant/nih/10704492. Licensed CC0.

---

*[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*