Building a Computational Model of Auditory Hallucinations

Psychotic disorders, including schizophrenia and bipolar disorder with psychotic features, are severe mental health conditions that significantly diminish quality of life and place a substantial economic burden on society. In England, the total societal cost of schizophrenia is estimated at £11.8 billion per year, with the public sector bearing approximately £7.2 billion of this cost. People with lived experience of psychosis (PWLEP) may experience a range of symptoms, including positive symptoms such as hallucinations and delusions, negative symptoms such as affective flattening, and cognitive impairments, including difficulties with attention, memory and executive function.

Auditory hallucinations, a common and disruptive symptom of psychotic disorders, are typically experienced as hearing voices and occur in approximately 75% of PWLEP. These voices are often intrusive and may be critical, threatening, or commanding, and can be associated with an increased risk of suicide. Whilst antipsychotic treatments, which primarily act through dopamine receptor antagonism, can be effective, they are often associated with significant side effects, and the response to antipsychotic treatment is insufficient in around 40% of patients. There is therefore an urgent need to better understand the mechanisms underlying auditory hallucinations to identify more effective treatment targets and develop novel therapies that act through mechanisms other than dopamine receptor antagonism.

A key avenue for understanding auditory hallucinations is examining disruptions in excitatory-inhibitory (E/I) balance, a fundamental mechanism in neural processing. Post-mortem and in vivo electrophysiological studies implicate dysregulated cortical microcircuits in psychosis, where excitatory pyramidal cells (signalling using glutamate) and inhibitory interneurons (signalling using gamma-aminobutyric acid or 'GABA') regulate neural activity. Interactions between these cells generate synchronised neural oscillations across different frequency bands, from low (delta, theta, alpha) to high (beta, gamma) frequencies. These oscillations, which are crucial for coordinating neural activity and supporting cognitive and perceptual function, can also serve as a measure of E/I balance and be assessed in humans using electroencephalography (EEG).

Previous studies have shown that resting-state cortical theta and gamma oscillations are increased, whilst beta oscillations and evoked gamma are reduced in people with psychosis, indicating altered E/I balance in the cortex. Dynamic causal modelling (DCM) is a validated computational technique for inferring E/I imbalance in both rodent models and humans, allowing a more precise estimate of E/I balance in the human brain to be obtained. Applying DCM to EEG and functional magnetic resonance imaging (fMRI) paradigms-including resting-state recordings, mismatch negativity (MMN), and the 40-Hz auditory steady-state response (ASSR)-has previously shown that pyramidal neuron disinhibition in the auditory cortex (primary auditory cortex and related regions) was associated with auditory perceptual symptoms, serving as a measure of hallucinations.

These findings have been demonstrated in those with chronic psychosis. The investigators now propose to measure E/I early in the presentation of a psychotic disorder to show whether this pathophysiology is related to symptoms demonstrated close to the onset of illness.

The aim of this study is to create a computational model of auditory hallucinations using cortical E/I balance in individuals with a psychotic disorder.

Primary Objective:

The principal research objective is to understand differences in brain signalling, specifically in E/I balance in those with auditory hallucinations using a computational model based on EEG measures such as ASSR relative to healthy volunteers without auditory hallucinations, and whether such measures predict changes in auditory hallucination over time.

Secondary Objectives:

The secondary research objectives include testing if levels of brain and blood substances and genetics are related to brain E/I balance in those with auditory hallucinations and healthy controls.

Inclusion criteria

PWLEP:

• Age 16-65 years.
• A score of at least 3 on the state section of the Auditory Perceptual Trait and State scale (APTS).
• If using an antipsychotic drug, participants must be on a stable dose of any medication during the baseline measures.
• Sufficient understanding of the nature of the study and any hazards of participating in it, or suitable consultee who can understand this and provide an opinion of what the participant's wishes would be if the participant does not have capacity, if appropriate to do so.
• Ability to communicate satisfactorily with the investigator and to participate in, and comply with the requirements of, the entire study.
• Capacity to give written consent to participate after reading the information and consent form and after having the opportunity to discuss the study with the investigators or their delegates, or suitable consultee who can understand this and provide an opinion of what the participant's wishes would be if the participant does not have capacity, if appropriate to do so.

Healthy controls:

• Age 16-65 years.
• Sufficient understanding of the nature of the study and any hazards of participating in it, or suitable consultee who can understand this and provide an opinion of what the participant's wishes would be if the participant does not have capacity, if appropriate to do so.
• Ability to communicate satisfactorily with the investigator and to participate in, and comply with the requirements of, the entire study.
• Capacity to give written consent to participate after reading the information and consent form and after having the opportunity to discuss the study with the investigators or their delegates, or suitable consultee who can understand this and provide an opinion of what the participant's wishes would be if the participant does not have capacity, if appropriate to do so.

Exclusion criteria

PWLEP:

• Clinically relevant abnormal history, or laboratory values at the pre-study screening assessment or participation in other research studies that could interfere with the objectives of the study or the safety of the participant, as deemed significant for a study researcher.
• Impaired endocrine, cardiac, pulmonary, thyroid, haematological, hepatic, respiratory, neurological, immunological or renal function, or other major disease (e.g. cancer) deemed clinically significant at the time of the study.
• Recent history (i.e. in the last 12 months) of epilepsy or seizures.
• Homicidal ideation or intent, as judged by a researcher; suicidal ideation, with some intent to act, or suicidal behaviour, as judged by a researcher.
• History of drug or alcohol dependence (except for caffeine and nicotine and mild/moderate cannabis dependence) in the 3 months before study start.
• Likelihood that the participant will not comply with the requirements of the protocol.
• Contraindication to MRI, EEG or other study procedures (for example history of metal implants or metal pieces entering the body that cannot be removed safely, severe claustrophobia, allergy or skin sensitivity to EEG electrode gel).
• Any other condition or disability that would make participation in the study procedures not possible (for example severe vision or hearing impairment).

Healthy controls:

As above, and also:

• No relevant family history of schizophrenia and psychosis (as judged relevant by a study researcher).