SIRS 2025: Insights on Schizophrenia Research, From Rights Based Care to Digital Biomarkers

Recently, I had the pleasure of attending the Schizophrenia International Research Society (SIRS) 2025 Congress, which brings together scientists, clinicians, industry members, and policymakers in this field. One of my favourite aspects of this conference is the opportunity to take in the breadth of research areas and methodologies aimed at increasing our understanding of schizophrenia and improving clinical care, with presentations spanning biological mechanisms, social determinants of health, computational modeling, and novel interventions, to name just a few of the topics presented.

Some of this year’s conference highlights for me included the “Rights based approach to mental health care” keynote address by Dr. Vikram Patel in which he laid out the fundamental case for a human rights-based approach to mental healthcare and strategies to support this aim; and one of the symposium sessions chaired by Dr. Henry Cowan and Dr. Vijay Mittal on “Emerging Advances in Transdiagnostic Phenotyping From Large-Scale Studies of Clinical High Risk for Psychosis”, which showcased several examples of how research on transdiagnostic symptom presentation, risk factors, and clinical trajectories in participants at clinical high-risk for psychosis can provide a pathway to improved identification and personalized prevention of psychosis risk. To me, these presentations served as a reminder that although the development of effective treatments for schizophrenia is an urgent and essential goal for our field, achieving this goal will only take us so far without also ensuring that 1) access to quality mental healthcare is available to all, and 2) that we improve our ability to identify and reduce psychosis risk.

Closer to my current research focus, I noted that many of the sessions continued to recognize the important contribution of negative symptoms and cognitive impairment to functional disability in schizophrenia, as well as the ongoing challenges in developing and evaluating effective interventions that demonstrate reliable and sustained improvement of these symptoms, particularly when moving from earlier to later phase clinical trials.

I had the opportunity to present two posters on the assessment of these important clinical domains of schizophrenia.

Poster 1: Identifying robust speech-based markers of negative symptom severity in schizophrenia
In the first poster, in collaboration with Dr. Sunny Tang from Northwell Health, we set out to identify robust speech-based markers of negative symptom severity from our brief digital speech assessment tasks. We analyzed speech responses from a sample of 62 inpatient participants with schizophrenia who completed Picture Description and verbal Journaling tasks at inpatient admission and discharge visits as part of a larger study on longitudinal predictors of psychosis outcomes.

We identified speech features showing robust correlations with negative symptom severity and then examined their psychometric properties. We found that 4 speech features were significantly and consistently correlated with greater clinician-rated negative symptom severity (on the Schedule for the Assessment of Negative Symptoms) across speech tasks and visits: longer mean pause duration, greater proportion of unfilled pauses, lower phonation rate, and slower speech rate. Importantly, none of these features were associated with positive symptoms of schizophrenia (demonstrating discriminant validity), nor were they associated with overall psychiatric symptom severity (demonstrating specificity to negative symptoms).

Speech feature test-retest reliability when repeating the tasks within visits was found to be acceptable-to-good for all 4 features on the Picture Description task and for the latter 3 features on the Journaling task. Additionally, speech rate was consistently correlated with a separate measure of negative symptom severity (demonstrating convergent validity) across the tasks and visits.

These findings suggest that several speech features, particularly speech rate, can serve as valid and reliable markers of negative symptom severity. Speech-based markers can complement traditional symptom ratings in the context of clinical trials; for example, by providing an objective and low-burden assessment of negative symptoms.

Poster 2: A meta-analysis on the sensitivity of CANTAB assessments to cognitive impairment associated with schizophrenia
The second poster, led by our chief scientist Dr. Francesca Cormack, with Edward Millgate (formerly at Cambridge Cognition) and Dr. Cecilie Koldbaek Lemvigh at the Center for Neuropsychiatric Schizophrenia Research (CNSR), presented the results of an updated meta-analysis on the sensitivity of CANTAB® digital cognitive assessments to cognitive impairment associated with schizophrenia (CIAS). The analysis focused on 7 CANTAB tasks corresponding to the 7 domains of cognition identified by the Measurement and Treatment Research to Improve Cognition in Schizophrenia (MATRICS) consensus process. 54 eligible published studies were identified, with the combined sample of participants with schizophrenia ranging from 83 to 1,383 depending on the CANTAB task. Significant moderate-to-large cognitive performance differences were seen between participants with schizophrenia vs. healthy control participants on all 7 CANTAB tasks, with Hedge’s g effect sizes of impaired performance ranging from g = 0.42 for verbal learning and memory (Verbal Recognition Memory task) to g = 1.1 for attention and vigilance (Rapid Visual Information Processing task). However, analyses also indicated that effect sizes for individual tasks did not significantly differ from the global cognitive impairment observed across tasks and studies (g = 0.89), consistent with the documented presence of generalized cognitive impairment in schizophrenia (e.g., Harvey et al., 2016; Schaefer et al., 2013). The results of this meta-analysis confirm that CANTAB assessments are sensitive to CIAS and provide context to help interpret the size of treatment effects in placebo-controlled trials.

DISCOURSE in Psychosis satellite meeting
Another highlight was the DISCOURSE Consortium Satellite Meeting taking place the day after the SIRS Congress. DISCOURSE is a diverse international scientific consortium for research into thought, language, and communication disturbances in psychosis. I had the privilege of presenting some of our clinical validation work on speech markers of negative symptom severity in schizophrenia, the findings of which can be found in our recent preprint (Spilka et al., 2024).

The meeting also featured updates from the consortium working groups, including the stakeholders working group, which is developing a survey to gauge diverse stakeholders’ (health care users, clinicians, industry members) views and expectations about the use of digital speech and language assessment in mental healthcare. Obtaining this input will help to ensure that research on speech biomarkers prioritizes the areas that matter most and that are likely to be most beneficial to these diverse stakeholders.

Overall, the SIRS conference highlighted promising developments and novel avenues to address ongoing challenges in our understanding and treatment of schizophrenia, and I believe this is an exciting time for the field.