Phonetic And Behavioral Markers Of Autobiographical Memory In Depressive Traits And Disorders

In recent years,the steadily increased major depressive disorder(MDD)prevalence bring heavy mental burden and economic loss to the families and society.The traditional diagnosis methods of depression depend on the doctors' clinical experience and the patients' self-cooperation,and it risks a series of subjective factors.Therefore,it's urgent to explore objective indictors of depression detection.Autobiographical memory(AM)refers to the recalling of personally life events,and previous studies demonstrated that depressed patients exhibits overgenerality(OGM).By conducting experiments on healthy college students and depressive patients,this study examined the behavioral and phonetic features of depressive traits and depressive disorders in AMs processing,and explored speech markers of depression detection.The first study explored the behavioral and phonetic features of depressive trait in AM processing.A total of 55 healthy college students(27 males)participated in the study and completed the AMs extraction task.The present study collected AMs of five emotions(happy,sad,neutral,fear,anger)through the dictation of the subjects,recorded their speech signal simultaneously and extracted some speech features including Energy,Zero Crossing Rate,Entropy of energy,Spectral Spread,Spectral Centroid,Spectral Entropy,Spectral Flux,Spectral Roll-off,Speaking Volume and Interrupt Number.Level of depression was assessed using Beck Depression Inventory(BDI-II)and the Patient Health Questionnaire-9(PHQ-9),and level of anxiety was assessed by State-Trait Anxiety Inventory(STAI).Results showed both of happy and fear specific AMs were negatively correlated with depression,and the remoteness of retrieved happy memories was positively correlated with depression.Since there were no significant association between depression and the behavioral of AMs,depression affected the specificity and intensity level of AMs through the mediation of the state anxiety.This shows that the state anxiety may affect the processing of AMs.In terms of phonetics,results showed a positive correlation between depression and spectral centroid of happy,sad and anger.Spectral centroid,spectral spread and spectral roll-off of sadness are significantly bigger than those of anger,indicating that there are specific emotional differences in speech features.This study shows that depression traits affect AMs and speech features of certain emotions,and are closely associated with the depression servity,which helps to better understand the processing mechanism of AMs.The second study examined the behavioral and phonetic features of depression in AMs processing.A total of 26 depressed patients(6 males)completed the tasks.The AM extraction paradigm and questionnaires were consistent with the first study,and Hamilton Depression Scale was also used for an vital assessment tool.Resulted showed positive correlation with depression were found in sad,fear and anger specific AMs,which is consistent with the previous research findings and findings of the first study,and indicated that individuals have more general memory with depression increased.Spectral entropy and spectral spread in depressed group were significantly higher than the control group,and depression may affect the voice of patients.Our study also showed that the emotional intensity of AMs and age can make better prediction on depression level,and energy,entropy of energy,spectral entropy,spectral flux,speaking volume and age also can have good prediction on depression level,which may provide assistance in early diagnostic depression.In sum,based on the impairment of AMs in depression,this paper employed the classical autobiographical memory extraction paradigm,and found that depression had impact on AMs and speech features of specific emotions,which helps to better understand the processing mechanism of AMs.The second study found that there are significant differences in some speech features between the depressed group and the control group,which is useful to distinguish the depressed and healthy group.