Publicação
Normative model detects abnormal functional connectivity in psychiatric disorders
| dc.contributor.author | Oliveira-Saraiva, Duarte | |
| dc.contributor.author | Ferreira, Hugo Alexandre | |
| dc.date.accessioned | 2025-03-13T17:47:54Z | |
| dc.date.available | 2025-03-13T17:47:54Z | |
| dc.date.issued | 2023 | |
| dc.description.abstract | Introduction: The diagnosis of psychiatric disorders is mostly based on the clinical evaluation of the patient’s signs and symptoms. Deep learning binary-based classification models have been developed to improve the diagnosis but have not yet reached clinical practice, in part due to the heterogeneity of such disorders. Here, we propose a normative model based on autoencoders. Methods: We trained our autoencoder on resting-state functional magnetic resonance imaging (rs-fMRI) data from healthy controls. The model was then tested on schizophrenia (SCZ), bipolar disorder (BD), and attention-deficit hyperactivity disorder (ADHD) patients to estimate how each patient deviated from the norm and associate it with abnormal functional brain networks’ (FBNs) connectivity. Rs-fMRI data processing was conducted within the FMRIB Software Library (FSL), which included independent component analysis and dual regression. Pearson’s correlation coefficients between the extracted blood oxygen level-dependent (BOLD) time series of all FBNs were calculated, and a correlation matrix was generated for each subject. Results and discussion: We found that the functional connectivity related to the basal ganglia network seems to play an important role in the neuropathology of BD and SCZ, whereas in ADHD, its role is less evident. Moreover, the abnormal connectivity between the basal ganglia network and the language network is more specific to BD. The connectivity between the higher visual network and the right executive control and the connectivity between the anterior salience network and the precuneus networks are the most relevant in SCZ and ADHD, respectively. The results demonstrate that the proposed model could identify functional connectivity patterns that characterize different psychiatric disorders, in agreement with the literature. The abnormal connectivity patterns from the two independent SCZ groups of patients were similar, demonstrating that the presented normative model was also generalizable. However, the group-level differences did not withstand individual-level analysis implying that psychiatric disorders are highly heterogeneous. These findings suggest that a precision-based medical approach, focusing on each patient’s specific functional network changes may be more beneficial than the traditional group-based diagnostic classification. | pt_PT |
| dc.description.version | info:eu-repo/semantics/publishedVersion | pt_PT |
| dc.identifier.doi | 10.3389/fpsyt.2023.1068397 | pt_PT |
| dc.identifier.uri | http://hdl.handle.net/10400.5/99311 | |
| dc.language.iso | eng | pt_PT |
| dc.peerreviewed | yes | pt_PT |
| dc.relation | Fundação para a Ciência e a Tecnologia (FCT) grant number UIDB/00645/2020 | pt_PT |
| dc.relation | Fundação para a Ciência e a Tecnologia (FCT) grant number DSAIPA/DS/0065/2018 | pt_PT |
| dc.subject | normative model | pt_PT |
| dc.subject | functional brain network | pt_PT |
| dc.subject | deep learning | pt_PT |
| dc.subject | psychiatric disorders | pt_PT |
| dc.subject | functional connectivity | pt_PT |
| dc.title | Normative model detects abnormal functional connectivity in psychiatric disorders | pt_PT |
| dc.type | journal article | |
| dspace.entity.type | Publication | |
| oaire.citation.title | Frontiers in Psychiatry | pt_PT |
| oaire.citation.volume | 14 | pt_PT |
| rcaap.rights | openAccess | pt_PT |
| rcaap.type | article | pt_PT |