1
|
Shoeibi A, Ghassemi N, Khodatars M, Moridian P, Khosravi A, Zare A, Gorriz JM, Chale-Chale AH, Khadem A, Rajendra Acharya U. Automatic diagnosis of schizophrenia and attention deficit hyperactivity disorder in rs-fMRI modality using convolutional autoencoder model and interval type-2 fuzzy regression. Cogn Neurodyn 2023; 17:1501-1523. [PMID: 37974583 PMCID: PMC10640504 DOI: 10.1007/s11571-022-09897-w] [Citation(s) in RCA: 8] [Impact Index Per Article: 8.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/12/2022] [Revised: 09/23/2022] [Accepted: 10/04/2022] [Indexed: 11/13/2022] Open
Abstract
Nowadays, many people worldwide suffer from brain disorders, and their health is in danger. So far, numerous methods have been proposed for the diagnosis of Schizophrenia (SZ) and attention deficit hyperactivity disorder (ADHD), among which functional magnetic resonance imaging (fMRI) modalities are known as a popular method among physicians. This paper presents an SZ and ADHD intelligent detection method of resting-state fMRI (rs-fMRI) modality using a new deep learning method. The University of California Los Angeles dataset, which contains the rs-fMRI modalities of SZ and ADHD patients, has been used for experiments. The FMRIB software library toolbox first performed preprocessing on rs-fMRI data. Then, a convolutional Autoencoder model with the proposed number of layers is used to extract features from rs-fMRI data. In the classification step, a new fuzzy method called interval type-2 fuzzy regression (IT2FR) is introduced and then optimized by genetic algorithm, particle swarm optimization, and gray wolf optimization (GWO) techniques. Also, the results of IT2FR methods are compared with multilayer perceptron, k-nearest neighbors, support vector machine, random forest, and decision tree, and adaptive neuro-fuzzy inference system methods. The experiment results show that the IT2FR method with the GWO optimization algorithm has achieved satisfactory results compared to other classifier methods. Finally, the proposed classification technique was able to provide 72.71% accuracy.
Collapse
Affiliation(s)
- Afshin Shoeibi
- FPGA Lab, Faculty of Electrical Engineering, K. N. Toosi University of Technology, Tehran, Iran
| | - Navid Ghassemi
- Computer Engineering Department, Ferdowsi University of Mashhad, Mashhad, Iran
| | - Marjane Khodatars
- Department of Medical Engineering, Mashhad Branch, Islamic Azad University, Mashhad, Iran
| | - Parisa Moridian
- Faculty of Engineering, Science and Research Branch, Islamic Azad University, Tehran, Iran
| | - Abbas Khosravi
- Institute for Intelligent Systems Research and Innovation (IISRI), Deakin University, Geelong, Australia
| | - Assef Zare
- Faculty of Electrical Engineering, Gonabad Branch, Islamic Azad University, Gonabad, Iran
| | - Juan M. Gorriz
- Department of Signal Theory, Networking and Communications, Universidad de Granada, Granada, Spain
| | | | - Ali Khadem
- Faculty of Electrical Engineering, K. N. Toosi University of Technology, Tehran, Iran
| | - U. Rajendra Acharya
- Ngee Ann Polytechnic, Singapore, 599489 Singapore
- Department of Biomedical Informatics and Medical Engineering, Asia University, Taichung, Taiwan
- Department of Biomedical Engineering, School of Science and Technology, Singapore University of Social Sciences, Singapore, Singapore
| |
Collapse
|
2
|
Mahmood U, Fu Z, Calhoun V, Plis S. GLACIER: GLASS-BOX TRANSFORMER FOR INTERPRETABLE DYNAMIC NEUROIMAGING. PROCEEDINGS OF THE ... IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING. ICASSP (CONFERENCE) 2023; 2023:10.1109/icassp49357.2023.10097126. [PMID: 37266485 PMCID: PMC10231935 DOI: 10.1109/icassp49357.2023.10097126] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 06/03/2023]
Abstract
Deep learning models can perform as well or better than humans in many tasks, especially vision related. Almost exclusively, these models are used to perform classification or prediction. However, deep learning models are usually of black-box nature, and it is often difficult to interpret the model or the features. The lack of interpretability causes a restrain from applying deep learning to fields such as neuroimaging, where the results must be transparent, and interpretable. Therefore, we present a 'glass-box' deep learning model and apply it to the field of neuroimaging. Our model mixes spatial and temporal dimensions in succession to estimate dynamic connectivity between the brain's intrinsic networks. The interpretable connectivity matrices produced by our model result in beating state-of-the-art models on many tasks using multiple functional MRI datasets. More importantly, our model estimates task-based flexible connectivity matrices, unlike static methods such as Pearson's correlation coefficients.
Collapse
Affiliation(s)
- Usman Mahmood
- Tri-institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS), Georgia State University, Georgia Institute of Technology, Emory University, Atlanta, GA, USA
- Georgia State University, Department of Computer Science, Atlanta, GA, USA
| | - Zening Fu
- Tri-institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS), Georgia State University, Georgia Institute of Technology, Emory University, Atlanta, GA, USA
- Georgia State University, Department of Computer Science, Atlanta, GA, USA
| | - Vince Calhoun
- Tri-institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS), Georgia State University, Georgia Institute of Technology, Emory University, Atlanta, GA, USA
- Georgia Institute of Technology, Department of Electrical and Computer Engineering, Atlanta, GA, USA
| | - Sergey Plis
- Tri-institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS), Georgia State University, Georgia Institute of Technology, Emory University, Atlanta, GA, USA
- Georgia State University, Department of Computer Science, Atlanta, GA, USA
| |
Collapse
|
3
|
Ionescu TM, Grohs-Metz G, Hengerer B. Functional ultrasound detects frequency-specific acute and delayed S-ketamine effects in the healthy mouse brain. Front Neurosci 2023; 17:1177428. [PMID: 37266546 PMCID: PMC10229773 DOI: 10.3389/fnins.2023.1177428] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/01/2023] [Accepted: 04/21/2023] [Indexed: 06/03/2023] Open
Abstract
Introduction S-ketamine has received great interest due to both its antidepressant effects and its potential to induce psychosis when administered subchronically. However, no studies have investigated both its acute and delayed effects using in vivo small-animal imaging. Recently, functional ultrasound (fUS) has emerged as a powerful alternative to functional magnetic resonance imaging (fMRI), outperforming it in sensitivity and in spatiotemporal resolution. In this study, we employed fUS to thoroughly characterize acute and delayed S-ketamine effects on functional connectivity (FC) within the same cohort at slow frequency bands ranging from 0.01 to 1.25 Hz, previously reported to exhibit FC. Methods We acquired fUS in a total of 16 healthy C57/Bl6 mice split in two cohorts (n = 8 received saline, n = 8 S-ketamine). One day after the first scans, performed at rest, the mice received the first dose of S-ketamine during the second measurement, followed by four further doses administered every 2 days. First, we assessed FC reproducibility and reliability at baseline in six frequency bands. Then, we investigated the acute and delayed effects at day 1 after the first dose and at day 9, 1 day after the last dose, for all bands, resulting in a total of four fUS measurements for every mouse. Results We found reproducible (r > 0.9) and reliable (r > 0.9) group-average readouts in all frequency bands, only the 0.01-0.27 Hz band performing slightly worse. Acutely, S-ketamine induced strong FC increases in five of the six bands, peaking in the 0.073-0.2 Hz band. These increases comprised both cortical and subcortical brain areas, yet were of a transient nature, FC almost returning to baseline levels towards the end of the scan. Intriguingly, we observed robust corticostriatal FC decreases in the fastest band acquired (0.75 Hz-1.25 Hz). These changes persisted to a weaker extent after 1 day and at this timepoint they were accompanied by decreases in the other five bands as well. After 9 days, the decreases in the 0.75-1.25 Hz band were maintained, however no changes between cohorts could be detected in any other bands. Discussion In summary, the study reports that acute and delayed ketamine effects in mice are not only dissimilar but have different directionalities in most frequency bands. The complementary readouts of the employed frequency bands recommend the use of fUS for frequency-specific investigation of pharmacological effects on FC.
Collapse
|
4
|
Mahmood U, Fu Z, Ghosh S, Calhoun V, Plis S. Through the looking glass: Deep interpretable dynamic directed connectivity in resting fMRI. Neuroimage 2022; 264:119737. [PMID: 36356823 PMCID: PMC9844250 DOI: 10.1016/j.neuroimage.2022.119737] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/28/2022] [Revised: 11/01/2022] [Accepted: 11/06/2022] [Indexed: 11/09/2022] Open
Abstract
Brain network interactions are commonly assessed via functional (network) connectivity, captured as an undirected matrix of Pearson correlation coefficients. Functional connectivity can represent static and dynamic relations, but often these are modeled using a fixed choice for the data window Alternatively, deep learning models may flexibly learn various representations from the same data based on the model architecture and the training task. However, the representations produced by deep learning models are often difficult to interpret and require additional posthoc methods, e.g., saliency maps. In this work, we integrate the strengths of deep learning and functional connectivity methods while also mitigating their weaknesses. With interpretability in mind, we present a deep learning architecture that exposes a directed graph layer that represents what the model has learned about relevant brain connectivity. A surprising benefit of this architectural interpretability is significantly improved accuracy in discriminating controls and patients with schizophrenia, autism, and dementia, as well as age and gender prediction from functional MRI data. We also resolve the window size selection problem for dynamic directed connectivity estimation as we estimate windowing functions from the data, capturing what is needed to estimate the graph at each time-point. We demonstrate efficacy of our method in comparison with multiple existing models that focus on classification accuracy, unlike our interpretability-focused architecture. Using the same data but training different models on their own discriminative tasks we are able to estimate task-specific directed connectivity matrices for each subject. Results show that the proposed approach is also more robust to confounding factors compared to standard dynamic functional connectivity models. The dynamic patterns captured by our model are naturally interpretable since they highlight the intervals in the signal that are most important for the prediction. The proposed approach reveals that differences in connectivity among sensorimotor networks relative to default-mode networks are an important indicator of dementia and gender. Dysconnectivity between networks, specially sensorimotor and visual, is linked with schizophrenic patients, however schizophrenic patients show increased intra-network default-mode connectivity compared to healthy controls. Sensorimotor connectivity was important for both dementia and schizophrenia prediction, but schizophrenia is more related to dysconnectivity between networks whereas, dementia bio-markers were mostly intra-network connectivity.
Collapse
Affiliation(s)
- Usman Mahmood
- Tri-institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS), Georgia State University, Georgia Institute of Technology, Emory University, Atlanta, GA, USA; Georgia State University, Department of Computer Science, Atlanta, GA, USA.
| | - Zening Fu
- Tri-institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS), Georgia State University, Georgia Institute of Technology, Emory University, Atlanta, GA, USA; Georgia State University, Department of Computer Science, Atlanta, GA, USA
| | - Satrajit Ghosh
- McGovern Institute for Brain Research, Massachusetts Institute of Technology, Cambridge, MA USA; Department of Otolaryngology - Head and Neck Surgery, Harvard Medical School, Boston, MA USA
| | - Vince Calhoun
- Tri-institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS), Georgia State University, Georgia Institute of Technology, Emory University, Atlanta, GA, USA; Georgia State University, Department of Computer Science, Atlanta, GA, USA; Georgia Institute of Technology, Department of Electrical and Computer Engineering, Atlanta, GA, USA
| | - Sergey Plis
- Tri-institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS), Georgia State University, Georgia Institute of Technology, Emory University, Atlanta, GA, USA; Georgia State University, Department of Computer Science, Atlanta, GA, USA
| |
Collapse
|
5
|
Altered spatio-temporal state patterns for functional dynamics estimation in first-episode drug-naive major depression. Brain Imaging Behav 2022; 16:2744-2754. [PMID: 36333522 PMCID: PMC9638404 DOI: 10.1007/s11682-022-00739-1] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Accepted: 10/18/2022] [Indexed: 11/06/2022]
Abstract
Patients with major depressive disorder (MDD) display affective and cognitive impairments. Although MDD-associated abnormalities of brain function and structure have been explored in depth, the relationships between MDD and spatio-temporal large-scale functional networks have not been evaluated in large-sample datasets. We employed data from International Big-Data Center for Depression Research (IBCDR), and comparable 543 healthy controls (HC) and 314 first-episode drug-naive (FEDN) MDD patients were included. We used a multivariate pattern classification method to learn informative spatio-temporal functional states. Brain states of each participant were extracted for functional dynamic estimation using an independent component analysis. Then, a multi-kernel pattern classification method was developed to identify discriminative spatio-temporal states associated with FEDN MDD. Finally, statistical analysis was applied to intrinsic and clinical brain characteristics. Compared with HC, FEDN MDD patients exhibited altered spatio-temporal functional states of the default mode network (DMN), the salience network, a hub network (centered on the dorsolateral prefrontal cortex), and a relatively complex coupling network (visual, DMN, motor-somatosensory and subcortical networks). Multi-kernel classification models to distinguish patients from HC obtained areas under the receiver operating characteristic curves up to 0.80. Classification scores correlated with Hamilton Depression Rating Scale scores and age at MDD onset. FEDN MDD patients had multiple abnormal spatio-temporal functional states. Classification scores derived from these states were related to symptom severity. The assessment of spatio-temporal states may represent a powerful clinical and research tool to distinguish between neuropsychiatric patients and controls.
Collapse
|
6
|
Lei D, Qin K, Pinaya WHL, Young J, Van Amelsvoort T, Marcelis M, Donohoe G, Mothersill DO, Corvin A, Vieira S, Lui S, Scarpazza C, Arango C, Bullmore E, Gong Q, McGuire P, Mechelli A. Graph Convolutional Networks Reveal Network-Level Functional Dysconnectivity in Schizophrenia. Schizophr Bull 2022; 48:881-892. [PMID: 35569019 PMCID: PMC9212102 DOI: 10.1093/schbul/sbac047] [Citation(s) in RCA: 2] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 11/28/2022]
Abstract
BACKGROUND AND HYPOTHESIS Schizophrenia is increasingly understood as a disorder of brain dysconnectivity. Recently, graph-based approaches such as graph convolutional network (GCN) have been leveraged to explore complex pairwise similarities in imaging features among brain regions, which can reveal abstract and complex relationships within brain networks. STUDY DESIGN We used GCN to investigate topological abnormalities of functional brain networks in schizophrenia. Resting-state functional magnetic resonance imaging data were acquired from 505 individuals with schizophrenia and 907 controls across 6 sites. Whole-brain functional connectivity matrix was extracted for each individual. We examined the performance of GCN relative to support vector machine (SVM), extracted the most salient regions contributing to both classification models, investigated the topological profiles of identified salient regions, and explored correlation between nodal topological properties of each salient region and severity of symptom. STUDY RESULTS GCN enabled nominally higher classification accuracy (85.8%) compared with SVM (80.9%). Based on the saliency map, the most discriminative brain regions were located in a distributed network including striatal areas (ie, putamen, pallidum, and caudate) and the amygdala. Significant differences in the nodal efficiency of bilateral putamen and pallidum between patients and controls and its correlations with negative symptoms were detected in post hoc analysis. CONCLUSIONS The present study demonstrates that GCN allows classification of schizophrenia at the individual level with high accuracy, indicating a promising direction for detection of individual patients with schizophrenia. Functional topological deficits of striatal areas may represent a focal neural deficit of negative symptomatology in schizophrenia.
Collapse
Affiliation(s)
| | | | - Walter H L Pinaya
- Department of Biomedical Engineering, School of Biomedical Engineering & Imaging Sciences, King’s College London, London, UK
| | - Jonathan Young
- Department of Neuroimaging, Institute of Psychiatry, Psychology, and Neuroscience, King’s College London, London, UK
| | - Therese Van Amelsvoort
- Department of Psychiatry and Neuropsychology, School for Mental Health and Neuroscience, Maastricht University Medical Center, Maastricht, The Netherlands
| | - Machteld Marcelis
- Department of Psychiatry and Neuropsychology, School for Mental Health and Neuroscience, Maastricht University Medical Center, Maastricht, The Netherlands
- Mental Health Care Institute Eindhoven (GGzE), Eindhoven, The Netherlands
| | - Gary Donohoe
- School of Psychology & Center for Neuroimaging and Cognitive Genomics, NUI Galway University, Galway, Ireland
| | - David O Mothersill
- Psychology Department, School of Business, National College of Ireland, Dublin, Ireland
| | - Aiden Corvin
- Department of Psychiatry, School of Medicine, Trinity College Dublin, Dublin, Ireland
| | - Sandra Vieira
- Department of Psychosis Studies, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, London, UK
| | - Su Lui
- Huaxi MR Research Center (HMRRC), Department of Radiology, West China Hospital, Sichuan University, Chengdu, China
| | - Cristina Scarpazza
- Department of Psychosis Studies, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, London, UK
- Department of General Psychology, University of Padova, Padova, Italy
- Padova Neuroscience Centre, University of Padova, Padova, Italy
| | - Celso Arango
- Institute of Psychiatry and Mental Health, Department of Child and Adolescent Psychiatry, Hospital General Universitario Gregorio Marañon, School of Medicine, Universidad Complutense Madrid, IiSGM, CIBERSAM, Madrid, Spain
| | - Ed Bullmore
- Brain Mapping Unit, Department of Psychiatry, University of Cambridge, Cambridge, UK
| | - Qiyong Gong
- To whom correspondence should be addressed; Huaxi MR Research Center (HMRRC), West China Hospital of Sichuan University, No 37 Guo Xue Xiang, Chengdu, 610041, China; tel: 86-18980601593, fax: 028-85423503,
| | - Philip McGuire
- Department of Psychosis Studies, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, London, UK
| | - Andrea Mechelli
- Department of Psychosis Studies, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, London, UK
| |
Collapse
|
7
|
Thalamic connectivity system across psychiatric disorders: Current status and clinical implications. BIOLOGICAL PSYCHIATRY GLOBAL OPEN SCIENCE 2021; 2:332-340. [PMID: 36324665 PMCID: PMC9616255 DOI: 10.1016/j.bpsgos.2021.09.008] [Citation(s) in RCA: 18] [Impact Index Per Article: 6.0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/28/2021] [Revised: 09/23/2021] [Accepted: 09/25/2021] [Indexed: 12/20/2022] Open
Abstract
The thalamic connectivity system, with the thalamus as the central node, enables transmission of the brain’s neural computations via extensive connections to cortical, subcortical, and cerebellar regions. Emerging reports suggest deficits in this system across multiple psychiatric disorders, making it a unique network of high translational and transdiagnostic utility in mapping neural alterations that potentially contribute to symptoms and disturbances in psychiatric patients. However, despite considerable research effort, it is still debated how this system contributes to psychiatric disorders. This review characterizes current knowledge regarding thalamic connectivity system deficits in psychiatric disorders, including schizophrenia, bipolar disorder, major depressive disorder, and autism spectrum disorder, across multiple levels of the system. We identify the presence of common and distinct patterns of deficits in the thalamic connectivity system in major psychiatric disorders and assess their nature and characteristics. Specifically, this review assembles evidence for the hypotheses of 1) thalamic microstructure, particularly in the mediodorsal nucleus, as a state marker of psychosis; 2) thalamo-prefrontal connectivity as a trait marker of psychosis; and 3) thalamo-somatosensory/parietal connectivity as a possible marker of general psychiatric illness. Furthermore, possible mechanisms contributing to thalamocortical dysconnectivity are explored. We discuss current views on the contributions of cerebellar-thalamic connectivity to the thalamic connectivity system and propose future studies to examine its effects at multiple levels, from the molecular (e.g., glutamatergic) to the behavioral (e.g., cognition), to gain a deeper understanding of the mechanisms that underlie the disturbances observed in psychiatric disorders.
Collapse
|