Detection of major depressive disorder, bipolar disorder, schizophrenia and generalized anxiety disorder using vocal acoustic analysis and machine learning: an exploratory study

Espinola, Caroline Wanderley; Gomes, Juliana Carneiro; Pereira, Jessiane Mônica Silva; Santos, Wellington Pinheiro dos

doi:10.1007/s42600-022-00222-2

Cited by 13 publications

(7 citation statements)

References 84 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Receiver operating characteristic analyses are commonly used for the diagnosis of neuropsychiatric disorders such as schizophrenia or major depression ( 24 – 26 ). To be reliable, a diagnostic indicator must exhibit an ROC value of 0.6 or higher ( 27 ).…”

Section: Discussionmentioning

confidence: 99%

Abnormal hubs in global network as neuroimaging biomarker in right temporal lobe epilepsy at rest

et al. 2022

View full text Add to dashboard Cite

While abnormal neuroimaging features have been reported in patients suffering from right temporal lobe epilepsy (rTLE), the value of altered degree centrality (DC) as a diagnostic biomarker for rTLE has yet to be established. As such, the present study was designed to examine DC abnormalities in rTLE patients in order to gauge the diagnostic utility of these neuroimaging features. In total, 68 patients with rTLE and 73 healthy controls (HCs) participated in this study. Imaging data were analyzed using DC and receiver operating characteristic (ROC) methods. Ultimately, rTLE patients were found to exhibit reduced right caudate DC and increased left middle temporal gyrus, superior parietal gyrus, superior frontal gyrus, right precuneus, frontal gyrus Inferior gyrus, middle-superior frontal gyrus, and inferior parietal gyrus DC relative to HC. ROC analyses indicated that DC values in the right caudate nucleus could be used to differentiate between rTLE patients and HCs with a high degree of sensitivity and specificity. Together, these results thus suggest that rTLE is associated with abnormal DC values in the right caudate nucleus, underscoring the relevance of further studies of the underlying pathophysiology of this debilitating condition.

show abstract

Section: Discussionmentioning

confidence: 99%

Abnormal hubs in global network as neuroimaging biomarker in right temporal lobe epilepsy at rest

et al. 2022

View full text Add to dashboard Cite

show abstract

“…For example, Zhuozheng Wang et al used electroencephalography (EGG) and a 2D convolutional network for depression diagnosis and achieved accuracy of 92% [ 37 ]. In a multiclass classification of major depressive disorder, bipolar disorder, schizophrenia, and generalized anxiety disorder, Caroline Wanderley Espinola et al reached 75.27% accuracy using a random forest classifier with 300 trees [ 38 ]. Another relevant work using the same activity data used in this paper was published by Rohit Kumar Bondugula et al, who achieved 86.60% accuracy in a bidirectional recurrent neural network (BRNN) for the binary classification of schizophrenic patients and healthy controls [ 39 ].…”

Section: Methodsmentioning

confidence: 99%

Classification of Depressive and Schizophrenic Episodes Using Night-Time Motor Activity Signal

et al. 2022

View full text Add to dashboard Cite

Major depressive disorder (MDD) is the most recurrent mental illness globally, affecting approximately 5% of adults. Furthermore, according to the National Institute of Mental Health (NIMH) of the U.S., calculating an actual schizophrenia prevalence rate is challenging because of this illness’s underdiagnosis. Still, most current global metrics hover between 0.33% and 0.75%. Machine-learning scientists use data from diverse sources to analyze, classify, or predict to improve the psychiatric attention, diagnosis, and treatment of MDD, schizophrenia, and other psychiatric conditions. Motor activity data are gaining popularity in mental illness diagnosis assistance because they are a cost-effective and noninvasive method. In the knowledge discovery in databases (KDD) framework, a model to classify depressive and schizophrenic patients from healthy controls is constructed using accelerometer data. Taking advantage of the multiple sleep disorders caused by mental disorders, the main objective is to increase the model’s accuracy by employing only data from night-time activity. To compare the classification between the stages of the day and improve the accuracy of the classification, the total activity signal was cut into hourly time lapses and then grouped into subdatasets depending on the phases of the day: morning (06:00–11:59), afternoon (12:00–17:59), evening (18:00–23:59), and night (00:00–05:59). Random forest classifier (RFC) is the algorithm proposed for multiclass classification, and it uses accuracy, recall, precision, the Matthews correlation coefficient, and F1 score to measure its efficiency. The best model was night-featured data and RFC, with 98% accuracy for the classification of three classes. The effectiveness of this experiment leads to less monitoring time for patients, reducing stress and anxiety, producing more efficient models, using wearables, and increasing the amount of data.

show abstract

“…Speech analysis is an emerging field in mental health research that holds great potential for improving the diagnosis and treatment of various mental disorders Cummins et al, 2015;De Boer et al, 2021;Morgan et al, 2021;Teferra et al, 2022;Wanderley Espinola et al, 2022). By analyzing patterns in language use, such as the use of specific words or the rate of speech, researchers can gain insights into a patient's cognitive and emotional functioning, as well as neural development (Mota, Natália Bezerra, Weissheimer, Finger, Ribeiro, Malcorra, & Hübner, 2023).…”

Section: Introductionmentioning

confidence: 99%

“…The potential to assess speech on a large scale, and at repeated intervals, could help in the identification of early signs of mental health problems and track changes in performance over time that might be predictive or indicative of changes in mental health. Indeed, previous work has shown exceptional promise of speech measures in terms of predicting, identifying and monitoring psychotic disorders such as schizophrenia or bipolar disorder (Agurto et al, Jul 2020a;Bedi et al, 2015;Pawełczyk et al, 2021;Wanderley Espinola et al, 2022).…”

Section: Introductionmentioning

confidence: 99%

Exploring the relationship between frequent cannabis use, vocal changes and language use

Olah,

Chesney,

Oliver

et al. 2023

Preprint

View full text Add to dashboard Cite

Background Cannabis use is prevalent among individuals with psychosis and a significant risk factor for the development of psychotic disorders, leading to unfavorable outcomes. Speech changes are frequently examined in psychotic disorders and connected to clinical outcomes. However, limited research investigated the impact of cannabis use on speech, which could potentially be a crucial confounding factor in the observed speech changes associated with psychosis. This study explores the differences in speech between frequent cannabis users and matched non-users. Methods Eight-minute speech samples from frequent cannabis users (N=82) and non-users (N=82) matched for gender, age, and education were manually transcribed and analyzed. Differences in vocal, semantic, and speech connectivity features were examined using T-tests. Logistic regression and random forest classifiers were employed to evaluate the discriminatory power of speech features. Results Acoustic, semantic and connectivity measures of speech showed significant differences between groups. Frequent cannabis users exhibited more voice breaks and back vowels. Connectivity analysis revealed more one-word loops in the cannabis group. Frequent users used more words, spoke coherently, and stayed on topic. Using logistic regression and random forest classifiers, participants were classified into cannabis or non-user groups with 0.66 / 0.72 accuracy (AUC = 0.62 / 0.69) based on demographic information and speech features. Conclusions Cannabis use is associated with alterations in speech features that overlap with those observed in psychosis. Consequently, understanding and potentially accounting for the effects of cannabis use on speech would be crucial for facilitating the treatment of psychotic disorders through speech analysis.

show abstract

Detection of major depressive disorder, bipolar disorder, schizophrenia and generalized anxiety disorder using vocal acoustic analysis and machine learning: an exploratory study

Cited by 13 publications

References 84 publications

Abnormal hubs in global network as neuroimaging biomarker in right temporal lobe epilepsy at rest

Abnormal hubs in global network as neuroimaging biomarker in right temporal lobe epilepsy at rest

Classification of Depressive and Schizophrenic Episodes Using Night-Time Motor Activity Signal

Exploring the relationship between frequent cannabis use, vocal changes and language use

Contact Info

Product

Resources

About