UncoveringSchizophrenia from verbal working memory tasks: An fnirs study
| Autor: | Amiyangshu De, Amit Konar, Souvik Biswas |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Recall
Working memory 05 social sciences 050109 social psychology Cognition medicine.disease behavioral disciplines and activities 050105 experimental psychology Task (project management) Statistical classification Schizophrenia medicine 0501 psychology and cognitive sciences Verbal memory Prefrontal cortex Psychology Cognitive psychology |
| Zdroj: | 2017 2nd International Conference for Convergence in Technology (I2CT). |
| DOI: | 10.1109/i2ct.2017.8226255 |
| Popis: | Schizophrenia is a mental disorder which has underlying neurological deficits. Schizophrenic patients have low social acceptance as well as a higher death rate. Initiation of schizophrenic symptoms are associated with a wide range of cognitive deficits. Impaired working memory performance is a basic characteristics of understanding schizophrenia. The performance impairment increases significantly with increased cognitive demand. Impaired verbal working memory is also associated with schizophrenic condition which can be measuredby analyzing prefrontal hemodynamics during various task difficulties. The brain regions associated with working memory performance is related to the prefrontal cortex. Therefore, near infrared spectroscopy becomes a significant tool for recording hemodynamic changes during working memory tasks. Our goal in this experiment is to classify schizophrenic and normal participants using three different task difficulties of verbal working memory and recall using differentconventional classification algorithms (LSVM, LDA and kNN). A principal component analysis induced feature selector-classifier based approach is undertaken to improve classification accuracy. LSVM provides the maximum classification accuracy of 85.12% for backward span verbal memory tasks. It is followed by LDA and KNN. The classifiers' performances are correlated with the performance accuracy data obtained from recall tasks. Among three different verbal working memory tasks, backward span recall is found to be the most appropriate in classifying schizophrenic from normal population. This paper reports a novel verbal working memory task performance based approach to diagnose schizophrenia. The experimental outcome supports the above mentioned possibility by providing a reliable classification accuracy in detecting schizophrenia from the normal population. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|