Explainable Depression Detection via Head Motion Patterns

Monika Gahalawat,Raul Fernandez Rojas,Tanaya Guha,Ramanathan Subramanian,Roland Goecke

2023-07-23

Abstract:While depression has been studied via multimodal non-verbal behavioural cues, head motion behaviour has not received much attention as a biomarker. This study demonstrates the utility of fundamental head-motion units, termed \emph{kinemes}, for depression detection by adopting two distinct approaches, and employing distinctive features: (a) discovering kinemes from head motion data corresponding to both depressed patients and healthy controls, and (b) learning kineme patterns only from healthy controls, and computing statistics derived from reconstruction errors for both the patient and control classes. Employing machine learning methods, we evaluate depression classification performance on the \emph{BlackDog} and \emph{AVEC2013} datasets. Our findings indicate that: (1) head motion patterns are effective biomarkers for detecting depressive symptoms, and (2) explanatory kineme patterns consistent with prior findings can be observed for the two classes. Overall, we achieve peak F1 scores of 0.79 and 0.82, respectively, over BlackDog and AVEC2013 for binary classification over episodic \emph{thin-slices}, and a peak F1 of 0.72 over videos for AVEC2013.

Computer Vision and Pattern Recognition,Machine Learning

What problem does this paper attempt to address?

The paper aims to address the issue of depression detection, specifically by identifying depression patients through head movement patterns. Specifically, the paper explores the following points:

1. **Research Background and Motivation**:
- Depression is a common mental health condition and has been recognized as one of the major contributors to the global health burden.
- Traditionally, the diagnosis of depression mainly relies on self-reports and clinical observations, which have certain subjectivity and bias.

2. **Research Objectives**:
- Explore the effectiveness of head movement as a biomarker for depression.
- Propose a method based on head movement units (referred to as "kinemes") for interpreting and detecting depression.
- Compare two methods: one that discovers kinemes from both depression patients and healthy controls; the other discovers kinemes only from healthy controls and calculates reconstruction errors to distinguish between patients and controls.

3. **Main Contributions**:
- For the first time, propose the use of kinemes as biomarkers for depression.
- Demonstrate that head movement patterns can serve as effective behavioral cues for detecting depression.
- Identify interpretative kinemes patterns consistent with previous studies.

4. **Experimental Results**:
- Experiments were conducted on the BlackDog and AVEC2013 datasets, achieving peak F1 scores of 0.79 and 0.82, respectively.
- Results indicate that kinemes discovered only from healthy controls perform better in distinguishing between patients and controls.

In summary, this paper proposes a new method for detecting depression by exploring head movement patterns and their kinemes representation, demonstrating the effectiveness and interpretability of this approach.