Final Report Summary - CODIP (Cortical Disorganization in Psychosis. Neurophysiological, clinical and genetic factors)
- Background, objectives and methods -
Schizophrenia and bipolar disorder are severe mental illnesses which cause high rates of disability, social and economic costs. Although there is compelling evidence of a genetic contribution to psychosis and we know deficits in neuronal connectivity are a key factor, the pathophysiology and aetiology of the disease remain poorly understood.
This project aimed: 1) To identify biomarkers of neural (dys-)connectivity characterising psychosis through neurophysiological and cognitive techniques, and 2) to apply such biomarkers in genetic association studies in order to understand the mechanisms by which variation at certain loci leads to the onset of the disease.
Neurophysiology is a non-invasive and inexpensive technique with a superb temporal resolution that allows the study of perceptual and cognitive processes in vivo. Our team has previously collected a large sample of families with psychosis cases including patients, unaffected relatives and healthy controls. Participants completed a battery of experiments including resting EEG, P300 wave, P50 gating, Mismatch Negativity and cognitive assessments. UCL, a centre of excellence in neuroscience, hosted this research. Dynamic Causal Modeling (DCM), which tests models of neural activity and its synchronisation deficits and is a new tool in psychiatry, and related candidate-gene association analyses were proposed to fulfil these objectives.
- Work performed -
The first part of this project was subdivided in two independent studies (Study #1: “Cortical (dys-)organization in psychosis - A classical quantitative EEG study” and Study #2: “Characterizing neural connectivity in psychosis using Dynamic Causal Modelling”). According the original project objectives, EEG data was successfully pre-processed, followed by time-frequency and connectivity analyses, as well as by a broad literature review. The researcher was trained in advanced programming (Matlab), signal processing software (EEGLAB, Fieldtrip and SPM 12b) and cutting-edge connectivity analysis methods (DCM). The latter spent most of the time during this project due its importance, complexity and the training required. As stated in the original project plan, and after a deep literature review, several hypothetical connectivity models (explicative of psychosis cortical disorganization) were selected and analysed using DCM.
This whole process required extensive training in Matlab programming, as well as deep DCM theoretical understanding and implementation. The fellow attended several training courses (e.g. the DCM course at the UCL Institute of Neurology); and closely collaborated with expert colleagues such as Prof. Karl Friston, Prof. Gareth Barnes, Dr. Dimitris Pinotsis and Dr. Rick Adams at the UCL Institute of Cognitive Neuroscience and the Wellcome Trust Centre for Neuroimaging. The fellow also attended these groups’ weekly meetings and presented his most recent results at different UCL departments, this way obtaining other researchers’ advice for improving his methodology and results interpretation. The newly acquired EEG analytic skills significantly improved the quality of fellow’s outcome, which derived in eight new publications (three as first author).
These two studies needed a few extra months to be concluded. This was required to guarantee results quality and reliability, and for increasing the options of being published in relevant scientific journals, as required in at this stage of the fellow’s postdoctoral training. The outcome of these DCM studies was presented in an international congress, internally at UCL meetings and symposiums, and in three original manuscripts currently under preparation and submission for publication.
Regarding the second part of this project (the genetic study), this was planned to be developed mainly during the second year of the project. As proposed in the original project, 1) genetic database development and preparation was successfully achieved, and 2) the fellow performed an extensive literature review and specialized training in R programming and the use of gene ontology tools to perform extensive genome-wide genotyping analyses. Genetic analyses (polygenetic risk scores) were developed over our neurophysiological, neuroanatomical and cognitive data on patients with psychosis, unaffected relatives and healthy controls.
- Main scientific results -
P300 and Mismatch Negativity (MMN), two evoked potential generated by auditory oddball stimuli that are significantly reduced in psychosis, were calculated and analysed with traditional EEG and Event-Related Potential (ERP) techniques. We measured changes in neuronal effective connectivity underlying these evoked potentials by using DCM. We hypothesised abnormal synaptic gain control in both individuals with psychosis and (to a lesser extent) in their first degree relatives.
DCM was used to estimate the excitability or self-inhibition of (source-specific) superficial neural activity in the three groups. P300 and MMN potentials were found altered in patients with psychosis and their unaffected relatives. The sources of the evoked potentials were estimated through previous DCM studies of neuroanatomical models and previous literature. Relevant sources were auditory primary cortices (Heschl’s gyri), superior temporal gyri, superior parietal loci and/or superior frontal gyri. Both patients with psychosis and their unaffected relatives (to a lesser degree) showed decreased activity but increased self-inhibition in frontal regions across task conditions, compared to controls. These results suggest that psychosis and genetic risk are associated with both context-dependent (condition-specific) and context-independent abnormalities of the excitability of superficial neural populations related to the studied related potentials. These abnormalities could relate to NMDA (neurotransmitter) receptor hypofunction on both pyramidal cells and inhibitory interneurons, and appear to be linked to the genetic aetiology of the illness, thereby constituting a potential endophenotype for psychosis. Hence, the next step was to study these potential biomarkers of psychoses in relation to the vulnerability genetics. Accordingly, we then calculated polygenic risk scores for schizophrenia and bipolar disorder using the latest mega-analyses from the Psychiatric Genomics Consortium, and tested whether this predicted performance on any of our neurophysiological, neuroanatomical and cognitive endophenotypes. After correction for multiple testing, none of the associations remained significant.
- Conclusion -
Through DCM we identified novel measures of brain (dys-)connectivity to use as phenotypes for genetic association. Biomarkers of brain connectivity could be used towards the earlier detection and treatment of psychosis. Understanding the mechanisms of action of susceptibility genes is essential to develop safer and more effective new treatments. Future research in line with this work may eventually develop better genetic-targeted clinical tools, thus improving patient’s outcomes and reducing related personal and economic costs. Attending the original research project, we consider that the main objectives were completely fulfilled. In June 2015, the researcher was awarded a prestigious Spanish ‘Juan de la Cierva’ fellowship to continue his research at the Centre for Biomedical Technology - Complutense University of Madrid.
Schizophrenia and bipolar disorder are severe mental illnesses which cause high rates of disability, social and economic costs. Although there is compelling evidence of a genetic contribution to psychosis and we know deficits in neuronal connectivity are a key factor, the pathophysiology and aetiology of the disease remain poorly understood.
This project aimed: 1) To identify biomarkers of neural (dys-)connectivity characterising psychosis through neurophysiological and cognitive techniques, and 2) to apply such biomarkers in genetic association studies in order to understand the mechanisms by which variation at certain loci leads to the onset of the disease.
Neurophysiology is a non-invasive and inexpensive technique with a superb temporal resolution that allows the study of perceptual and cognitive processes in vivo. Our team has previously collected a large sample of families with psychosis cases including patients, unaffected relatives and healthy controls. Participants completed a battery of experiments including resting EEG, P300 wave, P50 gating, Mismatch Negativity and cognitive assessments. UCL, a centre of excellence in neuroscience, hosted this research. Dynamic Causal Modeling (DCM), which tests models of neural activity and its synchronisation deficits and is a new tool in psychiatry, and related candidate-gene association analyses were proposed to fulfil these objectives.
- Work performed -
The first part of this project was subdivided in two independent studies (Study #1: “Cortical (dys-)organization in psychosis - A classical quantitative EEG study” and Study #2: “Characterizing neural connectivity in psychosis using Dynamic Causal Modelling”). According the original project objectives, EEG data was successfully pre-processed, followed by time-frequency and connectivity analyses, as well as by a broad literature review. The researcher was trained in advanced programming (Matlab), signal processing software (EEGLAB, Fieldtrip and SPM 12b) and cutting-edge connectivity analysis methods (DCM). The latter spent most of the time during this project due its importance, complexity and the training required. As stated in the original project plan, and after a deep literature review, several hypothetical connectivity models (explicative of psychosis cortical disorganization) were selected and analysed using DCM.
This whole process required extensive training in Matlab programming, as well as deep DCM theoretical understanding and implementation. The fellow attended several training courses (e.g. the DCM course at the UCL Institute of Neurology); and closely collaborated with expert colleagues such as Prof. Karl Friston, Prof. Gareth Barnes, Dr. Dimitris Pinotsis and Dr. Rick Adams at the UCL Institute of Cognitive Neuroscience and the Wellcome Trust Centre for Neuroimaging. The fellow also attended these groups’ weekly meetings and presented his most recent results at different UCL departments, this way obtaining other researchers’ advice for improving his methodology and results interpretation. The newly acquired EEG analytic skills significantly improved the quality of fellow’s outcome, which derived in eight new publications (three as first author).
These two studies needed a few extra months to be concluded. This was required to guarantee results quality and reliability, and for increasing the options of being published in relevant scientific journals, as required in at this stage of the fellow’s postdoctoral training. The outcome of these DCM studies was presented in an international congress, internally at UCL meetings and symposiums, and in three original manuscripts currently under preparation and submission for publication.
Regarding the second part of this project (the genetic study), this was planned to be developed mainly during the second year of the project. As proposed in the original project, 1) genetic database development and preparation was successfully achieved, and 2) the fellow performed an extensive literature review and specialized training in R programming and the use of gene ontology tools to perform extensive genome-wide genotyping analyses. Genetic analyses (polygenetic risk scores) were developed over our neurophysiological, neuroanatomical and cognitive data on patients with psychosis, unaffected relatives and healthy controls.
- Main scientific results -
P300 and Mismatch Negativity (MMN), two evoked potential generated by auditory oddball stimuli that are significantly reduced in psychosis, were calculated and analysed with traditional EEG and Event-Related Potential (ERP) techniques. We measured changes in neuronal effective connectivity underlying these evoked potentials by using DCM. We hypothesised abnormal synaptic gain control in both individuals with psychosis and (to a lesser extent) in their first degree relatives.
DCM was used to estimate the excitability or self-inhibition of (source-specific) superficial neural activity in the three groups. P300 and MMN potentials were found altered in patients with psychosis and their unaffected relatives. The sources of the evoked potentials were estimated through previous DCM studies of neuroanatomical models and previous literature. Relevant sources were auditory primary cortices (Heschl’s gyri), superior temporal gyri, superior parietal loci and/or superior frontal gyri. Both patients with psychosis and their unaffected relatives (to a lesser degree) showed decreased activity but increased self-inhibition in frontal regions across task conditions, compared to controls. These results suggest that psychosis and genetic risk are associated with both context-dependent (condition-specific) and context-independent abnormalities of the excitability of superficial neural populations related to the studied related potentials. These abnormalities could relate to NMDA (neurotransmitter) receptor hypofunction on both pyramidal cells and inhibitory interneurons, and appear to be linked to the genetic aetiology of the illness, thereby constituting a potential endophenotype for psychosis. Hence, the next step was to study these potential biomarkers of psychoses in relation to the vulnerability genetics. Accordingly, we then calculated polygenic risk scores for schizophrenia and bipolar disorder using the latest mega-analyses from the Psychiatric Genomics Consortium, and tested whether this predicted performance on any of our neurophysiological, neuroanatomical and cognitive endophenotypes. After correction for multiple testing, none of the associations remained significant.
- Conclusion -
Through DCM we identified novel measures of brain (dys-)connectivity to use as phenotypes for genetic association. Biomarkers of brain connectivity could be used towards the earlier detection and treatment of psychosis. Understanding the mechanisms of action of susceptibility genes is essential to develop safer and more effective new treatments. Future research in line with this work may eventually develop better genetic-targeted clinical tools, thus improving patient’s outcomes and reducing related personal and economic costs. Attending the original research project, we consider that the main objectives were completely fulfilled. In June 2015, the researcher was awarded a prestigious Spanish ‘Juan de la Cierva’ fellowship to continue his research at the Centre for Biomedical Technology - Complutense University of Madrid.