Indian Journal of PsychiatryIndian Journal of Psychiatry
Home | About us | Current Issue | Archives | Ahead of Print | Submission | Instructions | Subscribe | Advertise | Contact | Login 
    Users online: 1993 Small font sizeDefault font sizeIncrease font size Print this article Email this article Bookmark this page


    Advanced search

    Similar in PUBMED
    Email Alert *
    Add to My List *
* Registration required (free)  

    Artificial Intel...
    Cognitive functi...
    Exploring Therap...

 Article Access Statistics
    PDF Downloaded80    
    Comments [Add]    

Recommend this journal


 Table of Contents    
Year : 2020  |  Volume : 62  |  Issue : 7  |  Page : 10
Marfatia Awards

Click here for correspondence address and email

Date of Web Publication14-Jan-2020

How to cite this article:
. Marfatia Awards. Indian J Psychiatry 2020;62, Suppl S1:10

How to cite this URL:
. Marfatia Awards. Indian J Psychiatry [serial online] 2020 [cited 2021 Jun 20];62, Suppl S1:10. Available from:

   Artificial Intelligence-based classification of schizophrenia: A high density EEG- support vector machine study Top

Sai Krishna Tikka

Department of Psychiatry, All India Institute of Medical Sciences, Raipur

Bikesh Kumar Singh

Department of Bio-Medical Engineering,National Institute of Technology, Raipur Prof. S Haque Nizamie

Retired Professor of Psychiatry, Ranchi

Shobit Garg

Department of Psychiatry,

Shri Guru Ram Rai Institute of Medical & Health Sciences, Dehradun

Sunandan Mandal

Ph D Scholar, School of studies in Electronics & Photonics,

Pt. Ravishankar Shukla University, Raipur

Kavita Thakur

Professor, School of studies in Electronics & Photonics,

Pt. Ravishankar Shukla University, Raipur

Background: Interview-based schizophrenia diagnostic methods are not completely valid. Schizophrenia, the disease entity, is very heterogenous. Machine-Learning (ML) application of Artificial-Intelligence (AI) holds a tremendous promise in solving these issues.

Aims: To assess ML-based discriminating validity of resting state EEG quantitative features in classifying schizophrenia from healthy and positive and negative symptom subgroups, using high density recording.

Settings and Design: Data collected at a tertiary care mental hospital using a cross-sectional study design. Data analyzed at a premier Engineering Institute.

Methods and Material: Data of 38-schizophrenia patients and 20-healthy controls retrieved. The positive-negative subgroup classification done using PANSS operational-criteria. EEG recorded using 256-channel high density equipment. Eight priori regions-of-interest (ROIs) selected. Six-level, Daubechies-9 mother wavelet decomposition, Kernel-Support Vector Machine (SVM) method used for feature extraction and data classification.

Statistical analysis: Independent samples t-test and chi-square test for comparison of demographic and clinical variables. Mann-Whitney test for comparison of ML-features. Accuracy, sensitivity, specificity and area under receiver-operating-curve measured as discriminatory indices of classifications.

Results: Accuracy of classifying schizophrenia from healthy controls and, positive from negative symptom schizophrenia was 78.95% and 89.29%, respectively. While beta and gamma frequency related features most accurately classified schizophrenia from healthy controls, delta and theta frequency related features most accurately classified positive from negative schizophrenia. Inferior frontal gyrus related features most accurately contributed to both the classificatory instances.

Conclusions: SVM based classification and sub-classification of schizophrenia using EEG data is optimal and might help in improving the 'validity' and reducing the 'heterogeneity' in the diagnosis of schizophrenia.

Keywords: Machine-Learning; Positive symptoms; Feature-extraction; Negative symptoms; Validity.

   Cognitive function in Obsessive Compulsive Disorder and its correlation with BDNF Top

Debadatta Mohapatra, Department of Psychiatry, IMS & SUM Hospital, Bhubaneswar

Amrit Pattojoshi, Department of Psychiatry, Hi Tech Medical College, Bhubaneswar

Objective: In Obsessive Compulsive Disorder, serum brain-derived neurotrophic factor (BDNF) level decreases leading to dysfunctions of cellular plasticity and neuroprotective processes but at the same time some studies negate the role of BDNF in OCD. The present study was conducted to evaluate the change in serum BDNF level with cognitive function in OCD patients.

Methods: The present study is a prospective, interventional, clinical study conducted on 42 patients of obsessive compulsive disorder and 42 healthy controls. Detailed history, clinical evaluation, cognitive assessment with Montreal Cognitive Assessment was done. Serum BDNF was assessed for all 84 subjects.

Results: Various cognitive functions like visuospatial ability, executive function, attention, concentration were found to be significantly low in patients of OCD The serum BDNF level in the patients were compared with healthy controls and results revealed that there is a significant reduction ( p =0.002) in serum BDNF level in study group.

Conclusion: In OCD serum BDNF level is low and it is found to be directly related to the cognitive function.

KEY WORDS: Obsessive compulsive disorder; Brain-derived neurotrophic factor; Cognitive function.

   Exploring Therapeutic Effectiveness of a Novel Smartphone Application for Enhancing Cognitive Control: A Pilot Study in Obsessive Compulsive Disorder Top

Bappaditya Chowdhury, Department of Psychiatry,Institute of Neuroscience, Kolkata

Nilanjana Dutta Roy, Department of Computer Science And Engineering, Institute Of Engineering And Management, Salt-Lake, Kolkata

Introduction: In Research Domain Criteria (RDoC) project National Institute of Mental Health (NIMH) recommends Cognitive Control as a cross-cutting construct across diagnostic categories and advocates translational research for diagnostic and therapeutic approaches. Here we explore the therapeutic effectiveness of a novel Smart-phone application for enhancing cognitive control in Obsessive Compulsive Disorder as a model for pilot study.

Methodology: Two well recognized experimental paradigm for cognitive control, Switching Stroop Task and Stop Signal Reaction Time, were developed as an android based Smart-phone application. Individually tailored, disorder related symptom provoking images were incorporated in the application. A controlled application was also developed that was made free from the cognitive conflict present in the selected tasks. 21 subjects fulfilling study criteria were randomly assigned for a home-based practice with either therapeutic application or control application. Individuals in case and control group were assessed for adherence to task assigned, disorder related clinical outcome (YBOCS, WHOQOL-BREF, CGI/GI) and performance on assigned tasks (Switching Cost, Post Conflict Adaptation, Stop Signal Reaction Time).

Result: Subjects assigned with the therapeutic application showed more adherence to study (less drop-out p<0.00) more engagement in home-based practice as reflected by practice frequency (p<0.000). Improvement in disease related clinical parameter was noted in subjects assigned with therapeutic application (YBOCS p<0.000, WHOQOL-BREF p<0.02, CGI/GI p<0.01 ). There was some improvement related to neuropsychological tasks both in case and control group. This may be due to practice effect.

Conclusion: More stringent study is required for conclusive evidence of therapeutic effectiveness. However more engagement for subjects in cognitively challenging task (with cognitive conflict) was surprising. Probably some element of perceived therapeutic effectiveness was there. This may encourage further exploration with wider perspective.

Correspondence Address:
Login to access the Email id

Source of Support: None, Conflict of Interest: None

Rights and PermissionsRights and Permissions