Transdiagnostic Phenotyping Reveals a Host of Metacognitive Deficits Implicated in Compulsivity.
Tricia X F SeowClaire M GillanPublished in: Scientific reports (2020)
Recent work suggests that obsessive-compulsive disorder (OCD) patients have a breakdown in the relationship between explicit beliefs (i.e. confidence about states) and updates to behaviour. The precise computations underlying this disconnection are unclear because case-control and transdiagnostic studies yield conflicting results. Here, a large online population sample (N = 437) completed a predictive inference task previously studied in the context of OCD. We tested if confidence, and its relationship to action and environmental evidence, were specifically associated with self-reported OCD symptoms or common to an array of psychiatric phenomena. We then investigated if a transdiagnostic approach would reveal a stronger and more specific match between metacognitive deficits and clinical phenotypes. Consistent with prior case-control work, we found that decreases in action-confidence coupling were associated with OCD symptoms, but also 5/8 of the other clinical phenotypes tested (8/8 with no correction applied). This non-specific pattern was explained by a single transdiagnostic symptom dimension characterized by compulsivity that was linked to inflated confidence and several deficits in utilizing evidence to update confidence. These data highlight the importance of metacognitive deficits for our understanding of compulsivity and underscore how transdiagnostic methods may prove a more powerful alternative over studies examining single disorders.
Keyphrases
- case control
- obsessive compulsive disorder
- traumatic brain injury
- deep brain stimulation
- end stage renal disease
- newly diagnosed
- high throughput
- chronic kidney disease
- ejection fraction
- single cell
- high resolution
- prognostic factors
- genome wide
- patient reported
- peritoneal dialysis
- gene expression
- patient reported outcomes
- big data
- machine learning
- artificial intelligence
- mass spectrometry
- deep learning