Hugo G. Schnack
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Hugo G. Schnack has a total of 375 co-authors. Here are the top co-authors by number of shared publications:
- Hilleke E. Hulshoff Pol
Department of Psychiatry, University Medical Center Utrecht, Utrecht, The Netherlands
13 shared publications - René Sylvain Kahn
Department of Psychiatry, Icahn School of Medicine at Mount Sinai, New York, USA
9 shared publications Rachel M. Brouwer
Department of Psychiatry, UMC Brain Center, University Medical Center Utrecht, Utrecht, the Netherlands
7 shared publications- Elizabeth E. L. Buimer
Department of Psychiatry, UMCU Brain Center, University Medical Center Utrecht, Utrecht, The Netherlands
5 shared publications - Neeltje E. M. van Haren
Department of Psychiatry, University Medical Center Utrecht, Utrecht, The Netherlands
5 shared publications - Wiepke Cahn
Department of Psychiatry, University Medical Center Utrecht, Utrecht, The Netherlands
5 shared publications
Latest Publications
- Open Access
Prognostic predictions in psychosis: exploring the complementary role of machine learning models
BACKGROUND Predicting outcomes in schizophrenia spectrum disorders is challenging due to the variability of individual trajectories. While machine learning (ML) shows promise in outcome prediction, is has not yet been integrated into clinical practice. Understanding how ML models (MLMs) can complement psychiatrists’ predictions and bridge the gap between MLM capabilities and practical use is key. OBJECTIVE This study aims to compare the performance of psychiatrists and MLMs in predicting short-term symptomatic and functional remission in patients with first-episode psychosis and explore whether MLMs can improve psychiatrists’ prognostic accuracy. METHOD Twenty-four psychiatrists predicted symptomatic and functional remission probabilities based on written baseline information from 66 patients in the OPTiMiSE trial. ML-generated predictions were then shared with psychiatrists, allowing them to adjust their estimates. A questionnaire assessed trust in MLMs, perceived information gaps, and psychiatrists’ self-assessed predictive accuracy, which was compared to actual accuracy. FINDINGS The predictive accuracy of the MLM was comparable to that of psychiatrists for symptomatic remission (MLM: 0.50, psychiatrists: 0.52) and functional remission (MLM: 0.72, psychiatrists: 0.79). Interrater agreement was low but comparable for psychiatrists and the MLM. Although the MLM did not improve overall predictive accuracy, it showed potential in aiding psychiatrists with difficult-to-predict cases. However, psychiatrists struggled to recognize when to rely on the model’s output and we were unable to determine a clear pattern in these cases based on their characteristics. Psychiatrists could not reliably estimate their predictive accuracy. Psychiatrists expressed moderate to high trust in MLMs for prognostic prediction, but highlighted concerns about the lack of transparency and interpretability of model outputs. CONCLUSIONS MLMs are a promising tool for supporting psychiatric decision-making, particularly in challenging cases. However, their potential remains underutilized due to limitations in predictive accuracy and a lack of clarity in how predictions are generated. Addressing these issues is essential to build trust and foster integration into clinical practice. CLINICAL IMPLICATIONS MLMs are best suited as supplementary tools, providing a second opinion while psychiatrists retain decision-making autonomy. Integrating predictions from both sources may help reduce individual biases and improve accuracy. This approach leverages the strengths of MLMs without compromising clinical responsibility. What is already known on this topic While machine learning models (MLMs) show promise in predicting outcomes in psychotic disorders, they have yet to be integrated into clinical practice. Evidence on the predictive accuracy of psychiatrists for these disorders is limited, with only two small studies published before 1990 suggesting moderate accuracy. Comparisons of MLMs and psychiatrists in this context have not been previously conducted. What this study adds This is the first study to compare the predictive accuracy of psychiatrists with that of an MLM for psychotic disorders and to assess whether an MLM can enhance psychiatrists’ performance. It highlights that while MLMs do not improve overall accuracy, they may support psychiatrists in difficult cases. Insights into psychiatrists’ trust in MLMs and the challenges of implementing these models are also provided. How this study might affect research, practice, or policy The findings emphasize the need for advancements in MLM accuracy, interpretability, and strategies to identify cases where MLMs are most beneficial. These improvements could foster effective integration of MLMs as supplementary tools in clinical practice, aiding psychiatrists in decision-making while maintaining their autonomy.
Open AccessPrognostic predictions in psychosis: exploring the complementary role of machine learning models
BACKGROUND Predicting outcomes in schizophrenia spectrum disorders is challenging due to the variability of individual trajectories. While machine learning (ML) shows promise in outcome prediction, is has not yet been integrated into clinical practice. Understanding how ML models (MLMs) can complement psychiatrists’ predictions and bridge the gap between MLM capabilities and practical use is key. OBJECTIVE This study aims to compare the performance of psychiatrists and MLMs in predicting short-term symptomatic and functional remission in patients with first-episode psychosis and explore whether MLMs can improve psychiatrists’ prognostic accuracy. METHOD Twenty-four psychiatrists predicted symptomatic and functional remission probabilities based on written baseline information from 66 patients in the OPTiMiSE trial. ML-generated predictions were then shared with psychiatrists, allowing them to adjust their estimates. A questionnaire assessed trust in MLMs, perceived information gaps, and psychiatrists’ self-assessed predictive accuracy, which was compared to actual accuracy. FINDINGS The predictive accuracy of the MLM was comparable to that of psychiatrists for symptomatic remission (MLM: 0.50, psychiatrists: 0.52) and functional remission (MLM: 0.72, psychiatrists: 0.79). Interrater agreement was low but comparable for psychiatrists and the MLM. Although the MLM did not improve overall predictive accuracy, it showed potential in aiding psychiatrists with difficult-to-predict cases. However, psychiatrists struggled to recognize when to rely on the model’s output and we were unable to determine a clear pattern in these cases based on their characteristics. Psychiatrists could not reliably estimate their predictive accuracy. Psychiatrists expressed moderate to high trust in MLMs for prognostic prediction, but highlighted concerns about the lack of transparency and interpretability of model outputs. CONCLUSIONS MLMs are a promising tool for supporting psychiatric decision-making, particularly in challenging cases. However, their potential remains underutilized due to limitations in predictive accuracy and a lack of clarity in how predictions are generated. Addressing these issues is essential to build trust and foster integration into clinical practice. CLINICAL IMPLICATIONS MLMs are best suited as supplementary tools, providing a second opinion while psychiatrists retain decision-making autonomy. Integrating predictions from both sources may help reduce individual biases and improve accuracy. This approach leverages the strengths of MLMs without compromising clinical responsibility. What is already known on this topic While machine learning models (MLMs) show promise in predicting outcomes in psychotic disorders, they have yet to be integrated into clinical practice. Evidence on the predictive accuracy of psychiatrists for these disorders is limited, with only two small studies published before 1990 suggesting moderate accuracy. Comparisons of MLMs and psychiatrists in this context have not been previously conducted. What this study adds This is the first study to compare the predictive accuracy of psychiatrists with that of an MLM for psychotic disorders and to assess whether an MLM can enhance psychiatrists’ performance. It highlights that while MLMs do not improve overall accuracy, they may support psychiatrists in difficult cases. Insights into psychiatrists’ trust in MLMs and the challenges of implementing these models are also provided. How this study might affect research, practice, or policy The findings emphasize the need for advancements in MLM accuracy, interpretability, and strategies to identify cases where MLMs are most beneficial. These improvements could foster effective integration of MLMs as supplementary tools in clinical practice, aiding psychiatrists in decision-making while maintaining their autonomy.
- Open Access
- 10 December 2024
- Published by Springer Nature in Community Mental Health Journal
- Vol. 61 (2) , 300-313
In Pursuit of Recovery: A Comparative Study of Stakeholder Perspectives on Outcomes of People with Psychosis
There is no consensus on the definition of recovery of people with psychosis. This may be attributed to the broad concept including clinical, functional and personal recovery domains and the diverse interests of stakeholders involved. We hypothesized that service users, informal caregivers and healthcare professionals would exhibit differences in prioritization, reflecting varying roles and viewpoints. Through an open online anonymous questionnaire, stakeholders shared their perspectives on most important aspects of clinical, functional and personal recovery and on facilitators and barriers of recovery. Stakeholder collaboration in study design and interpretation enhanced the study’s robustness. The answers on open questions of respondents were categorized for statistical analyses to compare the answers between respondent groups on the different aspects of recovery. 226 stakeholders participated. While commonalities prevailed, distinctions emerged in recovery domain priorities. Service users and informal caregivers accorded equal importance to clinical, functional and personal recovery, whereas healthcare professionals emphasized clinical recovery. Regarding functional recovery, service users seemed to focus on practical issues while informal caregivers addressed the underlying causes. Disparities were notable in the role of healthcare that was considered as both a facilitator or a barrier to recovery, with informal caregivers prioritizing aspects of healthcare more frequently as most important than healthcare professionals. While commonalities prevailed, we identified several differences in prioritizing recovery domains among stakeholder groups. Awareness and understanding of these differences is crucial for effective communication and collaboration between stakeholders in recovery-oriented healthcare. The study underscores the need for ongoing dialogue between stakeholders.
- 10 December 2024
- Published by Springer Nature in Community Mental Health Journal
- Vol. 61 (2) , 300-313
Open AccessIn Pursuit of Recovery: A Comparative Study of Stakeholder Perspectives on Outcomes of People with Psychosis
There is no consensus on the definition of recovery of people with psychosis. This may be attributed to the broad concept including clinical, functional and personal recovery domains and the diverse interests of stakeholders involved. We hypothesized that service users, informal caregivers and healthcare professionals would exhibit differences in prioritization, reflecting varying roles and viewpoints. Through an open online anonymous questionnaire, stakeholders shared their perspectives on most important aspects of clinical, functional and personal recovery and on facilitators and barriers of recovery. Stakeholder collaboration in study design and interpretation enhanced the study’s robustness. The answers on open questions of respondents were categorized for statistical analyses to compare the answers between respondent groups on the different aspects of recovery. 226 stakeholders participated. While commonalities prevailed, distinctions emerged in recovery domain priorities. Service users and informal caregivers accorded equal importance to clinical, functional and personal recovery, whereas healthcare professionals emphasized clinical recovery. Regarding functional recovery, service users seemed to focus on practical issues while informal caregivers addressed the underlying causes. Disparities were notable in the role of healthcare that was considered as both a facilitator or a barrier to recovery, with informal caregivers prioritizing aspects of healthcare more frequently as most important than healthcare professionals. While commonalities prevailed, we identified several differences in prioritizing recovery domains among stakeholder groups. Awareness and understanding of these differences is crucial for effective communication and collaboration between stakeholders in recovery-oriented healthcare. The study underscores the need for ongoing dialogue between stakeholders.
- Open Access
- 26 September 2023
- Published by Cambridge University Press (CUP) in Psychological Medicine
- Vol. 54 (5) , 1016-1025
Accelerated brain aging as a biomarker for staging in bipolar disorder: an exploratory study
Background: Two established staging models outline the longitudinal progression in bipolar disorder (BD) based on episode recurrence or inter-episodic functioning. However, underlying neurobiological mechanisms and corresponding biomarkers remain unexplored. This study aimed to investigate if global and (sub)cortical brain structures, along with brain-predicted age difference (brain-PAD) reflect illness progression as conceptualized in these staging models, potentially identifying brain-PAD as a biomarker for BD staging.Methods: In total, 199 subjects with bipolar-I-disorder and 226 control subjects from the Dutch Bipolar Cohort with a high-quality T1-weighted magnetic resonance imaging scan were analyzed. Global and (sub)cortical brain measures and brain-PAD (the difference between biological and chronological age) were estimated. Associations between individual brain measures and the stages of both staging models were explored.Results: A higher brain-PAD (higher biological age than chronological age) correlated with an increased likelihood of being in a higher stage of the inter-episodic functioning model, but not in the model based on number of mood episodes. However, after correcting for the confounding factors lithium-use and comorbid anxiety, the association lost significance. Global and (sub)cortical brain measures showed no significant association with the stages.Conclusions: These results suggest that brain-PAD may be associated with illness progression as defined by impaired inter-episodic functioning. Nevertheless, the significance of this association changed after considering lithium-use and comorbid anxiety disorders. Further research is required to disentangle the intricate relationship between brain-PAD, illness stages, and lithium intake or anxiety disorders. This study provides a foundation for potentially using brain-PAD as a biomarker for illness progression.
- 26 September 2023
- Published by Cambridge University Press (CUP) in Psychological Medicine
- Vol. 54 (5) , 1016-1025
Open AccessAccelerated brain aging as a biomarker for staging in bipolar disorder: an exploratory study
Background: Two established staging models outline the longitudinal progression in bipolar disorder (BD) based on episode recurrence or inter-episodic functioning. However, underlying neurobiological mechanisms and corresponding biomarkers remain unexplored. This study aimed to investigate if global and (sub)cortical brain structures, along with brain-predicted age difference (brain-PAD) reflect illness progression as conceptualized in these staging models, potentially identifying brain-PAD as a biomarker for BD staging.Methods: In total, 199 subjects with bipolar-I-disorder and 226 control subjects from the Dutch Bipolar Cohort with a high-quality T1-weighted magnetic resonance imaging scan were analyzed. Global and (sub)cortical brain measures and brain-PAD (the difference between biological and chronological age) were estimated. Associations between individual brain measures and the stages of both staging models were explored.Results: A higher brain-PAD (higher biological age than chronological age) correlated with an increased likelihood of being in a higher stage of the inter-episodic functioning model, but not in the model based on number of mood episodes. However, after correcting for the confounding factors lithium-use and comorbid anxiety, the association lost significance. Global and (sub)cortical brain measures showed no significant association with the stages.Conclusions: These results suggest that brain-PAD may be associated with illness progression as defined by impaired inter-episodic functioning. Nevertheless, the significance of this association changed after considering lithium-use and comorbid anxiety disorders. Further research is required to disentangle the intricate relationship between brain-PAD, illness stages, and lithium intake or anxiety disorders. This study provides a foundation for potentially using brain-PAD as a biomarker for illness progression.
- Open Access
- 27 February 2023
- Published by Elsevier in Schizophrenia Research
- Vol. 254, 133-142
Systematic review and meta-analysis on predictors of prognosis in patients with schizophrenia spectrum disorders: An overview of current evidence and a call for prospective research and open access to datasets
- 27 February 2023
- Published by Elsevier in Schizophrenia Research
- Vol. 254, 133-142
Open AccessSystematic review and meta-analysis on predictors of prognosis in patients with schizophrenia spectrum disorders: An overview of current evidence and a call for prospective research and open access to datasets
- Top Cited Papers
- 5 April 2022
- Published by Springer Nature in Nature Neuroscience
- Vol. 25 (4) , 421-432
Genetic variants associated with longitudinal changes in brain structure across the lifespan
Human brain structure changes throughout the lifespan. Altered brain growth or rates of decline are implicated in a vast range of psychiatric, developmental and neurodegenerative diseases. In this study, we identified common genetic variants that affect rates of brain growth or atrophy in what is, to our knowledge, the first genome-wide association meta-analysis of changes in brain morphology across the lifespan. Longitudinal magnetic resonance imaging data from 15,640 individuals were used to compute rates of change for 15 brain structures. The most robustly identified genes GPR139, DACH1 and APOE are associated with metabolic processes. We demonstrate global genetic overlap with depression, schizophrenia, cognitive functioning, insomnia, height, body mass index and smoking. Gene set findings implicate both early brain development and neurodegenerative processes in the rates of brain changes. Identifying variants involved in structural brain changes may help to determine biological pathways underlying optimal and dysfunctional brain development and aging.
- 5 April 2022
- Published by Springer Nature in Nature Neuroscience
- Vol. 25 (4) , 421-432
Top Cited PapersGenetic variants associated with longitudinal changes in brain structure across the lifespan
Human brain structure changes throughout the lifespan. Altered brain growth or rates of decline are implicated in a vast range of psychiatric, developmental and neurodegenerative diseases. In this study, we identified common genetic variants that affect rates of brain growth or atrophy in what is, to our knowledge, the first genome-wide association meta-analysis of changes in brain morphology across the lifespan. Longitudinal magnetic resonance imaging data from 15,640 individuals were used to compute rates of change for 15 brain structures. The most robustly identified genes GPR139, DACH1 and APOE are associated with metabolic processes. We demonstrate global genetic overlap with depression, schizophrenia, cognitive functioning, insomnia, height, body mass index and smoking. Gene set findings implicate both early brain development and neurodegenerative processes in the rates of brain changes. Identifying variants involved in structural brain changes may help to determine biological pathways underlying optimal and dysfunctional brain development and aging.
- Open Access
- 22 March 2022
- Published by Society for Neuroscience in Journal of Neuroscience
- Vol. 42 (18) , 3704-3715
Longitudinal Allometry of Sulcal Morphology in Health and Schizophrenia
Scaling between subcomponents of folding and total brain volume (TBV) in healthy individuals (HIs) is allometric. It is unclear whether this is true in schizophrenia (SZ) or first-episode psychosis (FEP). This study confirmed normative allometric scaling norms in HIs using discovery and replication samples. Cross-sectional and longitudinal diagnostic differences in folding subcomponents were then assessed using an allometric framework. Structural imaging from a longitudinal (Sample 1: HI and SZ, nHI Baseline= 298, nSZ Baseline= 169, nHI Follow-up= 293, nSZ Follow-up= 168, totaling 1087 images, all individuals ≥ 2 images, age 16-69 years) and a cross-sectional sample (Sample 2: nHI= 61 and nFEP= 89, age 10-30 years), all human males and females, is leveraged to calculate global folding and its nested subcomponents: sulcation index (SI, total sulcal/cortical hull area) and determinants of sulcal area: sulcal length and sulcal depth. Scaling of SI, sulcal area, and sulcal length with TBV in SZ and FEP was allometric and did not differ from HIs. Longitudinal age trajectories demonstrated steeper loss of SI and sulcal area through adulthood in SZ. Longitudinal allometric analysis revealed that both annual change in SI and sulcal area was significantly stronger related to change in TBV in SZ compared with HIs. Our results detail the first evidence of the disproportionate contribution of changes in SI and sulcal area to TBV changes in SZ. Longitudinal allometric analysis of sulcal morphology provides deeper insight into lifespan trajectories of cortical folding in SZ.SIGNIFICANCE STATEMENTPsychotic disorders are associated with deficits in cortical folding and brain size, but we lack knowledge of how these two morphometric features are related. We leverage cross-sectional and longitudinal samples in which we decompose folding into a set of nested subcomponents: sulcal and hull area, and sulcal depth and length. We reveal that, in both schizophrenia and first-episode psychosis, (1) scaling of subcomponents with brain size is different from expected scaling laws and (2) caution is warranted when interpreting results from traditional methods for brain size correction. Longitudinal allometric scaling points to loss of sulcal area as a principal contributor to loss of brain size in schizophrenia. These findings advance the understanding of cortical folding atypicalities in psychotic disorders.
- 22 March 2022
- Published by Society for Neuroscience in Journal of Neuroscience
- Vol. 42 (18) , 3704-3715
Open AccessLongitudinal Allometry of Sulcal Morphology in Health and Schizophrenia
Scaling between subcomponents of folding and total brain volume (TBV) in healthy individuals (HIs) is allometric. It is unclear whether this is true in schizophrenia (SZ) or first-episode psychosis (FEP). This study confirmed normative allometric scaling norms in HIs using discovery and replication samples. Cross-sectional and longitudinal diagnostic differences in folding subcomponents were then assessed using an allometric framework. Structural imaging from a longitudinal (Sample 1: HI and SZ, nHI Baseline= 298, nSZ Baseline= 169, nHI Follow-up= 293, nSZ Follow-up= 168, totaling 1087 images, all individuals ≥ 2 images, age 16-69 years) and a cross-sectional sample (Sample 2: nHI= 61 and nFEP= 89, age 10-30 years), all human males and females, is leveraged to calculate global folding and its nested subcomponents: sulcation index (SI, total sulcal/cortical hull area) and determinants of sulcal area: sulcal length and sulcal depth. Scaling of SI, sulcal area, and sulcal length with TBV in SZ and FEP was allometric and did not differ from HIs. Longitudinal age trajectories demonstrated steeper loss of SI and sulcal area through adulthood in SZ. Longitudinal allometric analysis revealed that both annual change in SI and sulcal area was significantly stronger related to change in TBV in SZ compared with HIs. Our results detail the first evidence of the disproportionate contribution of changes in SI and sulcal area to TBV changes in SZ. Longitudinal allometric analysis of sulcal morphology provides deeper insight into lifespan trajectories of cortical folding in SZ.SIGNIFICANCE STATEMENTPsychotic disorders are associated with deficits in cortical folding and brain size, but we lack knowledge of how these two morphometric features are related. We leverage cross-sectional and longitudinal samples in which we decompose folding into a set of nested subcomponents: sulcal and hull area, and sulcal depth and length. We reveal that, in both schizophrenia and first-episode psychosis, (1) scaling of subcomponents with brain size is different from expected scaling laws and (2) caution is warranted when interpreting results from traditional methods for brain size correction. Longitudinal allometric scaling points to loss of sulcal area as a principal contributor to loss of brain size in schizophrenia. These findings advance the understanding of cortical folding atypicalities in psychotic disorders.
- Open Access
- 4 August 2021
- Published by Cambridge University Press (CUP) in Psychological Medicine
- Vol. 53 (4) , 1302-1312
Acoustic speech markers for schizophrenia-spectrum disorders: a diagnostic and symptom-recognition tool
Background: Clinicians routinely use impressions of speech as an element of mental status examination. In schizophrenia-spectrum disorders, descriptions of speech are used to assess the severity of psychotic symptoms. In the current study, we assessed the diagnostic value of acoustic speech parameters in schizophrenia-spectrum disorders, as well as its value in recognizing positive and negative symptoms.Methods: Speech was obtained from 142 patients with a schizophrenia-spectrum disorder and 142 matched controls during a semi-structured interview on neutral topics. Patients were categorized as having predominantly positive or negative symptoms using the Positive and Negative Syndrome Scale (PANSS). Acoustic parameters were extracted with OpenSMILE, employing the extended Geneva Acoustic Minimalistic Parameter Set, which includes standardized analyses of pitch (F0), speech quality and pauses. Speech parameters were fed into a random forest algorithm with leave-ten-out cross-validation to assess their value for a schizophrenia-spectrum diagnosis, and PANSS subtype recognition.Results: The machine-learning speech classifier attained an accuracy of 86.2% in classifying patients with a schizophrenia-spectrum disorder and controls on speech parameters alone. Patients with predominantly positive v. negative symptoms could be classified with an accuracy of 74.2%.Conclusions: Our results show that automatically extracted speech parameters can be used to accurately classify patients with a schizophrenia-spectrum disorder and healthy controls, as well as differentiate between patients with predominantly positive v. negatives symptoms. Thus, the field of speech technology has provided a standardized, powerful tool that has high potential for clinical applications in diagnosis and differentiation, given its ease of comparison and replication across samples.
- 4 August 2021
- Published by Cambridge University Press (CUP) in Psychological Medicine
- Vol. 53 (4) , 1302-1312
Open AccessAcoustic speech markers for schizophrenia-spectrum disorders: a diagnostic and symptom-recognition tool
Background: Clinicians routinely use impressions of speech as an element of mental status examination. In schizophrenia-spectrum disorders, descriptions of speech are used to assess the severity of psychotic symptoms. In the current study, we assessed the diagnostic value of acoustic speech parameters in schizophrenia-spectrum disorders, as well as its value in recognizing positive and negative symptoms.Methods: Speech was obtained from 142 patients with a schizophrenia-spectrum disorder and 142 matched controls during a semi-structured interview on neutral topics. Patients were categorized as having predominantly positive or negative symptoms using the Positive and Negative Syndrome Scale (PANSS). Acoustic parameters were extracted with OpenSMILE, employing the extended Geneva Acoustic Minimalistic Parameter Set, which includes standardized analyses of pitch (F0), speech quality and pauses. Speech parameters were fed into a random forest algorithm with leave-ten-out cross-validation to assess their value for a schizophrenia-spectrum diagnosis, and PANSS subtype recognition.Results: The machine-learning speech classifier attained an accuracy of 86.2% in classifying patients with a schizophrenia-spectrum disorder and controls on speech parameters alone. Patients with predominantly positive v. negative symptoms could be classified with an accuracy of 74.2%.Conclusions: Our results show that automatically extracted speech parameters can be used to accurately classify patients with a schizophrenia-spectrum disorder and healthy controls, as well as differentiate between patients with predominantly positive v. negatives symptoms. Thus, the field of speech technology has provided a standardized, powerful tool that has high potential for clinical applications in diagnosis and differentiation, given its ease of comparison and replication across samples.
- Open Access
- 2 July 2021
- Published by Springer Nature in Schizophrenia
- Vol. 7 (1) , 1-11
Individualized prediction of three- and six-year outcomes of psychosis in a longitudinal multicenter study: a machine learning approach
Schizophrenia and related disorders have heterogeneous outcomes. Individualized prediction of long-term outcomes may be helpful in improving treatment decisions. Utilizing extensive baseline data of 523 patients with a psychotic disorder and variable illness duration, we predicted symptomatic and global outcomes at 3-year and 6-year follow-ups. We classified outcomes as (1) symptomatic: in remission or not in remission, and (2) global outcome, using the Global Assessment of Functioning (GAF) scale, divided into good (GAF ≥ 65) and poor (GAF < 65). Aiming for a robust and interpretable prediction model, we employed a linear support vector machine and recursive feature elimination within a nested cross-validation design to obtain a lean set of predictors. Generalization to out-of-study samples was estimated using leave-one-site-out cross-validation. Prediction accuracies were above chance and ranged from 62.2% to 64.7% (symptomatic outcome), and 63.5–67.6% (global outcome). Leave-one-site-out cross-validation demonstrated the robustness of our models, with a minor drop in predictive accuracies of 2.3% on average. Important predictors included GAF scores, psychotic symptoms, quality of life, antipsychotics use, psychosocial needs, and depressive symptoms. These robust, albeit modestly accurate, long-term prognostic predictions based on lean predictor sets indicate the potential of machine learning models complementing clinical judgment and decision-making. Future model development may benefit from studies scoping patient’s and clinicians' needs in prognostication.
- 2 July 2021
- Published by Springer Nature in Schizophrenia
- Vol. 7 (1) , 1-11
Open AccessIndividualized prediction of three- and six-year outcomes of psychosis in a longitudinal multicenter study: a machine learning approach
Schizophrenia and related disorders have heterogeneous outcomes. Individualized prediction of long-term outcomes may be helpful in improving treatment decisions. Utilizing extensive baseline data of 523 patients with a psychotic disorder and variable illness duration, we predicted symptomatic and global outcomes at 3-year and 6-year follow-ups. We classified outcomes as (1) symptomatic: in remission or not in remission, and (2) global outcome, using the Global Assessment of Functioning (GAF) scale, divided into good (GAF ≥ 65) and poor (GAF < 65). Aiming for a robust and interpretable prediction model, we employed a linear support vector machine and recursive feature elimination within a nested cross-validation design to obtain a lean set of predictors. Generalization to out-of-study samples was estimated using leave-one-site-out cross-validation. Prediction accuracies were above chance and ranged from 62.2% to 64.7% (symptomatic outcome), and 63.5–67.6% (global outcome). Leave-one-site-out cross-validation demonstrated the robustness of our models, with a minor drop in predictive accuracies of 2.3% on average. Important predictors included GAF scores, psychotic symptoms, quality of life, antipsychotics use, psychosocial needs, and depressive symptoms. These robust, albeit modestly accurate, long-term prognostic predictions based on lean predictor sets indicate the potential of machine learning models complementing clinical judgment and decision-making. Future model development may benefit from studies scoping patient’s and clinicians' needs in prognostication.
- 1 May 2021
- Published by Elsevier in Biological Psychiatry
- Vol. 89 (9) , S371-S372
Patterns of Structural Covariance Abnormalities and Clinical Correlations in Schizophrenia
- 1 May 2021
- Published by Elsevier in Biological Psychiatry
- Vol. 89 (9) , S371-S372
Patterns of Structural Covariance Abnormalities and Clinical Correlations in Schizophrenia
- Open Access
- 1 May 2021
- Published by Elsevier in Schizophrenia Research
- Vol. 231, 189-197
Accelerated aging in the brain, epigenetic aging in blood, and polygenic risk for schizophrenia
- 1 May 2021
- Published by Elsevier in Schizophrenia Research
- Vol. 231, 189-197
Open AccessAccelerated aging in the brain, epigenetic aging in blood, and polygenic risk for schizophrenia