The melody of depression: Acoustic measures of depressed patients in short-term psychodynamic therapy

Aviv Nof & Sigal Zilcha-Mano

The aim of the acoustic project is to explore the benefits of acoustic features (markers) of patients as indicators for process and measured outcome of MDD in psychotherapy.
Previous empirical studies on MDD and acoustic features were conducted mostly in laboratory rather than therapeutic settings. These studies show promising results. For example, associations have been reported between reduced fundamental frequency mean and range and increased levels of depression severity (Honig et al., 2014), depressed people were reported to produce longer speech pauses (Carroll, 1981), and depression was characterized by decreases speech rate (Balsters et al., 2012).
We propose a novel approach, based on acoustic analysis to develop an objective, time-efficient assessment method that can be conducted accurately even by junior staff. Sixty patients with MDD, in a randomized controlled trial study, will be randomly assigned to two types of short-term psychodynamic therapy (supportive and supportive-expressive conditions). Each therapy session will be digitally recorded and acoustically analyzed. The acoustic analyses will be based on four markers, that were recommended to be used in psychotherapy research by Rochaman & Amir (2013): fundamental frequency, jitter, speaking rate, and the ratio of pauses (silences).
The proposed research has three objectives, to be achieved by means of three studies: Study 1 will focus on the acoustic markers as prognostic predictors of effective treatment; Study 2 will focus on changes in markers as indicators of progress throughout treatment; and Study 3 will focus on the synchronicity in acoustic markers between patient and therapist as a predictor of subsequent treatment outcome.
We believe that psychotherapy practice will change in the coming years and incorporate automatic computer analysis of markers and real-time evaluations with artificial intelligence procedures, to help psychologists interpret complicated processes, including evaluation procedures, progress analysis, and automatic detection of in-session problems. The current study has potential to contribute to this task with the acoustic analysis procedures.