The Relationship Between Speech Features Changes When You Get Depressed: Feature Correlations for Improving Speed and Performance of Depression Detection
This work addresses depression detection for mental health applications, but it is incremental as it builds on existing methods like SVMs and LSTMs with a new data representation.
The paper tackles depression detection from speech by showing that depression alters correlations between speech features, and using feature correlation matrices instead of vectors improves training speed and performance. Results include a relative error rate reduction of 23.1% to 26.6% on a dataset of 112 speakers.
This work shows that depression changes the correlation between features extracted from speech. Furthermore, it shows that using such an insight can improve the training speed and performance of depression detectors based on SVMs and LSTMs. The experiments were performed over the Androids Corpus, a publicly available dataset involving 112 speakers, including 58 people diagnosed with depression by professional psychiatrists. The results show that the models used in the experiments improve in terms of training speed and performance when fed with feature correlation matrices rather than with feature vectors. The relative reduction of the error rate ranges between 23.1% and 26.6% depending on the model. The probable explanation is that feature correlation matrices appear to be more variable in the case of depressed speakers. Correspondingly, such a phenomenon can be thought of as a depression marker.