Quantifying oxytocin effects on vocal expression in schizophrenia

点击翻译按钮获取中文摘要

Schizophrenia is a devastating illness associated with lifelong disability and high health care costs that disproportionately impacts veterans. Negative symptoms, a set of volitional and expressive deficits, are major contributors to impaired functioning. These deficits are poorly understood and difficult to monitor, in part due to a lack of effective measurement tools. Negative symptoms are typically measured using interview-based clinical rating scales, which are imprecise, costly to administer, and rely on behavior observed in constrained laboratory and clinical environments. Speech is a key indicator of clinical status and an easily collected resource that can be leveraged to address this gap. Abnormal speech is a hallmark of schizophrenia that reflects expressive deficits: patients tend to talk less and pause more while talking (i.e. alogia) and have decreased musicality and emotion in their voice (i.e. blunted vocal affect). Advances in automated analytic methods and mobile device capability provide an opportunity to dramatically improve quantification of speech abnormalities with unprecedented efficiency. Automated analysis of veterans’ speech, combined with remote speech data collection using mobile devices, can enable precise, frequent, and cost-effective measurement of negative symptoms across laboratory, clinical, and real-world settings. The ability to obtain rich, quantitative characterizations of negative symptoms at the individual level will serve to elucidate pathophysiology of specific deficits and transform our ability to monitor veterans’ clinical status, thus impacting both research and clinical care. This CDA-1 leverages already-collected laboratory data and adds novel mobile data collection methods to Dr. Josh Woolley’s Merit-funded clinical trial to generate preliminary data on the clinical relevance and feasibility of using automated methods to measure speech abnormalities in veterans with schizophrenia. The program aims to: (1) investigate how automatically quantified speech abnormalities relate to gold standard clinical ratings of negative symptoms and functioning in people with schizophrenia (n=50); (2) examine the potential of oxytocin (OT)—a candidate treatment for expressive deficits—to improve speech abnormalities in men with schizophrenia (n=30) who have already completed a randomized, placebo-controlled, cross-over trial; (3) pilot the collection of speech data (both recorded audio samples and passively-extracted vocal signals) outside the laboratory via mobile devices in veterans with schizophrenia (n=20); and (4) explore the links between functional neural connectivity, speech abnormalities, and clinically rated negative symptoms in veterans with schizophrenia (n=20) who will complete neuroimaging as part of the Merit trial. The training plan will focus on developing critical quantitative and logistical skills; specifically: (1) automated speech analysis using an established analytic approach; (2) remote speech data collection and processing via mobile devices using the mobile Ecological Momentary Assessment application; and (3) functional magnetic resonance imaging (fMRI) processing and resting-state functional connectivity (rsFC) analyses. These research and training aims will yield critical preliminary data and skills that lay the groundwork for a CDA-2 that will determine OT effects on speech abnormalities and their functional and neural correlates using automated analysis of speech data collected remotely throughout Dr. Woolley’s Merit trial. The proposed program is the first step towards a broader long-term goal: to develop scalable methods for high-resolution, low-cost quantification of deficits associated with serious neuropsychiatric illness that will deepen understanding of their functional and neural correlates, accelerate development of targeted treatments, and enhance efficient detection of changes in clinical status to improve health care for veterans. Automated speech analysis and remote data collection offer a promising route to this goal and have the potential to measure deficits associated with multiple neuropsychiatric disorders impacting veterans such as depression, traumatic brain injury, and Parkinson’s disease.