Table 2.
Particulars of publicly accessible datasets for SZ diagnosis.
| Dataset | Publisher | Modality | Number of Samples | Sample Diversity | Data Collection Biases | Download Link |
|---|---|---|---|---|---|---|
| COBRE [56]—2012 | Mind Research Network and the University of New Mexico | sMRI and fMRI | SZ = 72 HC = 75 |
Primarily U.S.-based, limited demographic diversity (age 18–65) | Demographic homogeneity, exclusion of comorbid conditions, institutional differences | http://fcon_1000.projects.nitrc.org/indi/retro/cobre.html (accessed on 20 July 2024) |
| RepOD [57]—2017 | Institute of Psychiatry and Neurology in Warsaw, Poland | EEG | SZ = 14 HC = 14 |
Focused on paranoid schizophrenia subtype | Exclusion of other psychiatric disorders | https://repod.icm.edu.pl/dataset.xhtml?persistentId=doi:10.18150/repod.0107441 (accessed on 15 September 2023) |
| NUSDAST [58]—2015 | National University of Singapore (NUS) | sMRI | SZ = 171 HC = 170 Strict SZ = 44 No disorder = 66 |
Mixed demographics | Potential institutional bias, overrepresentation of certain clinical subtypes | https://sites.wustl.edu/oasisbrains/central-xnat/ (accessed on 22 July 2024) |
| UCLA [59]—2013 | OpenfMRI project | sMRI and fMRI and DWI | SZ = 50 HC = 130 Patients with ADHD = 43 Bipolar illness = 49 |
Diverse group covering SZ, bipolar disorder, ADHD, and HC | Possible overrepresentation of specific disorders (SZ, bipolar), lack of detailed demographic data. | https://openfmri.org/dataset/ds000030/ (accessed on 22 July 2024) |
| SchizConnect [60]—2015 | SchizConnect consortium | sMRI and fMRI | No disorder = 632 Strict SZ = 384 Broad SZ = 215 Bipolar disorder = 10 Schizoaffective disorder = 41 Siblings of unknown disorder = 66 Siblings of strict SZ = 44 |
Diverse diagnostic groups, limited demographic details | Overrepresentation of certain diagnostic groups (strict SZ, broad SZ), potential selection bias from varied data sources | http://www.schizconnect.org/ (accessed on 22 July 2024) |
| MCIC [61]—2013 | Mind Research Network (MRN) | sMRI and fMRI and DWI | SZ = 162 HC = 169 |
Wide age range, limited to specific geographic and clinical settings | Potential bias due to geographic focus | https://www.nitrc.org/projects/mcic/ (accessed on 20 July 2024) |
| MLSP2014 [62]—2014 | IEEE | sMRI and fMRI | SZ = 69 HC = 75 |
Limited demographic variation | Potential bias from focus on structural and functional brain features | https://www.kaggle.com/c/mlsp-2014-mri (accessed on 20 July 2024) |
| MSU [63]—2017 | Lomonosov Moscow State University | EEG | SZ = 45 HC = 39 |
Specific to age group, limited broader representation | Age-specific focus (adolescents) | http://brain.bio.msu.ru/eeg_schizophrenia.htm (accessed on 15 September 2023) |
| Phase II of FBIRN [64]—2007 | FBIRN consortium | fMRI | SZ = 87 HC = 85 |
Moderate diversity, spanning ages 18–70 years, limited geographic representation | Potential overrepresentation of specific SZ subtypes, age-related bias | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4651841/ (accessed on 23 July 2024) |
| Phase III of FBIRN [64]—2007 | SZ = 176 HC = 186 |
Broader sample with similar age range (18–70 years), limited demographic diversity | Potential selection bias due to specific inclusion criteria |