Raghunathan, Karthik Charan (2022) Voicing a Schizophrenic Mind: A natural language processing based approach to diagnose schizophrenia using speech. Research Project 1 (minor thesis), Behavioural and Cognitive Neurosciences.
|
Text
mBCN_2022_KarthikCharanRaghunathan.pdf Download (5MB) | Preview |
|
![[img]](https://fse.studenttheses.ub.rug.nl/style/images/fileicons/text.png) |
Text
Toestemming.pdf Restricted to Registered users only Download (124kB) |
Abstract
Language disturbances are a prominent symptom of schizophrenia. Current studies involve analysing the speech of subjects to distinguish schizophrenic patients from healthy controls. In such studies, the machine learning (ML) models learn from single speech recordings of different subjects. This introduces noise in the models due to voice based differences among different individuals. . The present study proposes analysing the changes in speech of individual subjects over a period of time to predict whether the patient is transitioning into a relapse. Speech data of subjects undergoing consultation for schizophrenia were collected in the form of a semi-structured interview. Acoustic parameters were extracted from these speech recordings using OpenSMILE. These parameters were fed into different ML and deep learning models as a time series to compare their performance in classifying between a relapse vs non-relapse.In order to replicate the results from previous studies, speech data of the subjects was fed into a support vector machine (SVM) and random forest classifiers. To model the inter-individual differences in speech, the data, remodelled as a time series was fed into a 3-layered deep neural network and a convolutional neural network (CNN). Finally, the rate of change of the parameters were analysed.
| Item Type: | Thesis (Research Project 1 (minor thesis)) |
|---|---|
| Supervisor name: | Vugt, M.K. van |
| Degree programme: | Behavioural and Cognitive Neurosciences |
| Thesis type: | Research Project 1 (minor thesis) |
| Language: | English |
| Date Deposited: | 09 Aug 2022 06:52 |
| Last Modified: | 09 Aug 2022 06:52 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/28299 |
Actions (login required)
 |
View Item |