Active Mental Health

Improving Cognitive Health in Early Psychosis and Depression with AI assisted Cognitive Remediation

In plain English

AI plain-English summary

Cognitive difficulties—problems with memory, attention, and planning—often block recovery for people with psychosis and depression, and an AI chatbot is being developed to help deliver a proven therapy that tackles them. This matters because the therapy, called Cognitive Remediation (CR), works but is rarely used. The programme, CIRCuiTSTM, combines task practice, therapist coaching, and strategy teaching to improve thinking skills and daily functioning. Independent studies have shown it helps young people with psychosis, and evidence is growing for depression. But delivering it requires extensive therapist time, making it too expensive for most clinical settings. The AI chatbot is designed to take over some of that work, allowing the therapy to reach more patients without sacrificing quality. If successful, this could change how mental health services treat cognitive problems. Instead of offering CR only in specialist centres with dedicated therapists, the AI-supported version could be rolled out across the NHS and similar health systems. The research includes an observational study to provide the National Institute for Health and Care Excellence with the evidence needed for broader implementation. For patients, this could mean better thinking skills, greater independence, and a real chance at recovery—not just symptom control.

View original technical description
For individuals with mental health conditions, cognitive difficulties often stand as a major barrier to achieving recovery and independence. Over the last three decades, Cognitive Remediation (CR), an evidence-based psychological intervention, has been shown to improve both cognition and functioning. At King’s College London, we developed CIRCuiTSTM, an innovative CR programme that integrates key elements such as task practice, therapist support, strategy teaching, and support for generalisation, while also focusing on metacognition, a critical factor in accelerating learning and skills transfer. Independent studies confirmed the benefits of CIRCuiTSTM in young people with psychosis, with growing evidence supporting its efficacy in other conditions.Despite robust evidence for CR, global adoption remains limited due to the significant therapist time required, posing challenges to implement it in resource-constrained clinical settings. This research addresses this gap by developing an AI chatbot to support the safe, effective, and efficient delivery of CIRCuiTSTM. Co-produced with experts by experience and computer scientists, the AI-supported CR will be trialled with young people experiencing psychosis and depression. In partnership with industry, we will the conduct an observational study to provide the National Institute for Health and Care Excellence with evidence for broader implementation of CR in health services.

View the original record at the funder ↗

Researchers

Alexandra Kenny (EPMC Awardee)Charles Wells (EPMC Awardee)Dominic MacBean (EPMC Awardee)Huajie (Lily) Jin (EPMC Awardee)Matteo Cella (EPMC Awardee)Richard Emsley (EPMC Awardee)Rumina Taylor (EPMC Awardee)Stephen Ireland (EPMC Awardee)Til Wykes (EPMC Awardee)Tom Price (EPMC Awardee)

Related Research

Grants with similar aims, by meaning.

CIRCuiTS a novel cognitive remediation web-based intervention
Exploratory randomised controlled trial of computerised cognitive remediation therapy (CIRCuiTS) for people with schizophrenia
Evaluating remote delivery of Cognitive Remediation for people severe mental health conditions
Narrowing the Needs-Provision Gap for Young People's Mental Health: Developing and Piloting an AI-Guided Behavioural Activation Intervention
Optimising recovery following the first psychotic episode by enhancing cognitive reserve

Original classification

Mental Health Award: Accelerating scalable digital mental health interventions

Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.