Mild depression and normal sadness: Language diagnostic criteria

Objective: High rates of depression are viewed with some degree of skepticism due to the issue of validity and reliability of psychiatric diagnoses. The difficulties in the diagnostic differentiation of ordinary sadness and depression in the continuum of depressive states, and the existence of false-positive diagnoses of depression are identified. Incorrect diagnostics of mild depressions is related to clinical interpretation of thinking content and structure within the depressive triad. Whereas nonverbal (motor) and affective components are similar to variations of normal sadness in healthy individuals, associative component reveals the most sensitive psychopathological unit, which is represented in language changes.

Methods: The study sample of 124 patients aged 41.85±11.89 years (67% female) was taken from Russian native speakers diagnosed with mild depression (D) in primary care and consulted at the University Psychiatry Department. 77 healthy persons (65% female) (H), including 35 healthy persons with reactions of normal sadness (NS), were observed as controls. Speech was studied using a number of standard psycholinguistic procedures at the superficial and deep levels of language. Statistical data analysis included descriptive methodics, nonparametric analysis (U-criteria Mann-Whitney, test by Wald-Wolfowitz, p<0,05), mathematic modeling of discriminate analysis (λ–Wilks; method Standard).

Results: Lexical-stylistic sublevel of speech in D was characterized by verbosity (D-311.18±11.43 words in text; H-197.25±11.24), inversive word order (D-124(100%);H-5(6.49%)), narration (D-106(85.48%);H-56(72.73%)) dominated over reasoning (D-18 (14.52%);H-21(27.27%)), communicative discourse signs (P-124(100%);H-9(11.69%)), increased number of phraseologisms (P-3.74;H-1.21), tautologies (P-3.77;H-1.44), repetitions (P-4.42;H-1.82), inversions (P-4.00;H-1.08), ellipsis (P-3.73;H-1.60). Lexical-grammar sublevel contained more pronouns of all types with the prevalence of personal (P-124(100%);H-54(70.13%) pronouns, continuous form verbs (P-116(93.55%);H-26 (33.77%)) in past tense (P-124(100%);H-2(2.6%)). Syntaxical-stylistic sublevel represented the prevalence of simple (P-99(79.84%);H-2(2.6%)) sentences, truncated and impersonal (P-84(67.74%);H-7(9,.09%)) types. Component analysis of patients’ speech demonstrated the distortion of quality and reduction of semantic component (over evaluated existential and family categories, diminished cognitive and altruistic categories, overestimation of communicative and hedonic categories over the self-realization and social status). In accordance to clinical criteria, the sample of mild depression was subdivided onto subtypes (anxious, asthenic-hypodynamic, melancholic) where appropriate language differences were also revealed. Speech in NS, in comparison to H, was impoverished by ellipsis (NS-1.60, H-1.21), tautologies (NS-1.66,H-1.22) and lexical, semantic repetitions (NS-1.97,H-1.69). Tempo of speech was reduced due to defaults (NS-1.71,H-1.38). Persons used preferentially verbs in continuous form (NS-54.29%,H-16.67%) in present tense (NS-88.57%,H-47.62%). Mathematical modeling proved the hypothesis about importance of language criteria and significance of verbal markers for diagnostics and differentiation of mild depression and normal sadness (98%, λ–Wilks=0.0007;p<0.001). Conclusions: 1) Mild depression and state of normal sadness are characterized by definite psycholinguistic features, which reşect cognitive dysfunction and thought strategies significantly related to the symptoms of leading hypothymic affect and signs of mood deviation. 2) In mild depression and state of normal sadness, the speech is distorted both in structure and in semantics. 3) The structure of affective component significantly determines the whole mental status and is represented in speech, so the sensitivity of linguistic markers in diagnostics of mild depression and normal sadness is revealed.