Freely Generated Word Responses Analyzed With Artificial Intelligence Predict Self-Reported Symptoms of Depression, Anxiety, and Worry

Abstract

BackgroundQuestion-based computational language assessments (QCLA) of mental health, based on self-reported and freely generated word responses and analyzed with artificial intelligence, is a potential complement to rating scales for identifying mental health issues. This study aimed to examine to what extent this method captures items related to the primary and secondary symptoms associated with Major Depressive Disorder (MDD) and Generalized Anxiety Disorder (GAD) described in the Diagnostic and Statistical Manual of Mental Disorders (DSM-5). We investigated whether the word responses that participants generated contained information of all, or some, of the criteria that define MDD and GAD using symptom-based rating scales that are commonly used in clinical research and practices.MethodParticipants (N = 411) described their mental health with freely generated words and rating scales relating to depression and worry/anxiety. Word responses were quantified and analyzed using natural language processing and machine learning.ResultsThe QCLA correlated significantly with the individual items connected to the DSM-5 diagnostic criteria of MDD (PHQ-9; Pearson’s r = 0.30–0.60, p < 0.001) and GAD (GAD-7; Pearson’s r = 0.41–0.52, p < 0.001; PSWQ-8; Spearman’s r = 0.52–0.63, p < 0.001) for respective rating scales. Items measuring primary criteria (cognitive and emotional aspects) yielded higher predictability than secondary criteria (behavioral aspects).ConclusionTogether these results suggest that QCLA may be able to complement rating scales in measuring mental health in clinical settings. The approach carries the potential to personalize assessments and contributes to the ongoing discussion regarding the diagnostic heterogeneity of depression.

Dispositional and ideological factor correlate of conspiracy thinking and beliefs

“…These text samples are turned into semantic spaces through singular value decomposition (SVD), a technique akin to principal component analysis (PCA). Through this technique, it is possible to obtain stable numerical estimates on the likelihood of sequences of words used in the diagnosis of thought disorders [49] or how closely groups of words such as item texts or generated free texts assemble each other [60][61][62], for example how close a given text comes to a diagnostically recognized statement [47,63]. The numerical output of LSA is usually a cosine, where numbers approaching 1 indicate identical meaning between the compared texts, and lower numbers indicate disparate or unrelated meanings in the compared texts.…”

“…This type of thinking has shown itself to be detectable through digital text analysis [ 19 , 42 , 43 ], more specifically Latent Semantic Analysis (LSA). This technique has recently gained more interest as a method in psychology [ 44 – 46 ] as it allows a statistical comparison of cognitive content in texts, adding to the use of rating scales [ 47 , 48 ]. The thought patterns peculiar to schizophrenia and Cluster A conditions have been found to show up in statistically significant ways using LSA, predicting these conditions better than human judges [ 49 ].…”

“…The thought patterns peculiar to schizophrenia and Cluster A conditions have been found to show up in statistically significant ways using LSA, predicting these conditions better than human judges [ 49 ]. For this reason, we will subject the item contents to LSA to shed light on the response patterns [ 50 ], an approach akin to what has earlier been found to work the diagnostics of personality and psychopathology in the schizotypal spectrum [ 19 , 47 , 49 ]. If cognitive peculiarities are indeed central to a propensity for believing in CTs we expect that the semantics of the conspiracy theories are most closely related to the semantics of the items measuring schizotypal PDs (H9).…”

Assessment of Depression and Anxiety in Young and Old with a Question Based Computational Language Approach

“…The question-based computational language assessment method (QCLA) can be applied to semantic data (i.e. words and sentences), where the assessment is based on high-dimensional word embeddings from a large language corpus (Kjell et al, 2021). Kjell et al, (2021) investigated word response relating to the symptoms of major depressive disorder (MDD) and generalised anxiety disorders (GAD) as described in the Diagnostic and Statistical Manual of Mental Disorders (DSM-5).…”

Precise Language Responses Challenge Easy Rating Scales - Comparing Clinicians’ and Respondents’ Views

“…Freely generated word responses analysed with NLP have been found to correlate with the primary and secondary symptoms in DSM-5 that are associated with Major Depressive Disorder (MDD) and Generalized Anxiety Disorder (GAD; Kjell, Johnsson, Sikström, 2021). QCLA has also been shown to correlate with theoretically relevant cooperative behaviour, especially for prosocial individuals in a give-some dilemma game, whereas rating scales were unable to show the expected effect (Kjell, Daukantaitë & Sikström, 2021).…”