2D data arrangement to train ANN for depression levels measurement

Area of Science:

• Psychology
• Computer Science
• Health Informatics

Background:

• Depression significantly impacts physical health, but solely physical markers are insufficient for accurate measurement.
• Integrating psychological and physical data offers a more comprehensive approach to understanding depression.

Purpose of the Study:

• To develop an Artificial Neural Network (ANN) model for depression level measurement.
• To investigate the utility of two-dimensional (2D) data, combining physical and psychological parameters, for enhanced depression detection.

Main Methods:

• A dataset of 95 college students was collected, including four noninvasive physical parameters and Perceived Stress Scale (PSS) scores.
• Artificial Neural Networks (ANNs) were trained using this 2D dataset to identify patterns between physical and psychological indicators.

Main Results:

• Initial analysis indicated significant correlations between perceived stress and physical parameters like elevated heart rate and reduced sleep quality.
• A highly significant p-value supports the observed differences in means, suggesting non-coincidental findings.
• The developed dataset encompasses all depression levels, aiming to improve measurement effectiveness.

Conclusions:

• Combining physical and psychological data in a 2D approach enhances the accuracy of Artificial Neural Networks (ANNs) in detecting depression.
• This integrated data strategy is crucial for advancing depression measurement tools.