A Linguistic Analysis of Instagram Captions Between Adolescent Suicide Decedents and Living Controls

Background: Suicide rates continue to rise for adolescents in the United States. 62% of teenagers use Instagram, and as technology and research in this domain advance, social media posts could provide insights into near-term adolescent risk states and could inform new strategies for suicide prevention. This study analyzed language in captions of teenagers' Instagram accounts in the 3 months before suicide and compared caption language to matched living controls. Method: The study identified 89 teenagers who died by suicide using obituaries and news reports and 89 matched living control teenagers. Linguistic Inquiry and Word Count (LIWC) software was used to test for differences in specific language categories across linguistic, psychological, and topical categories (e.g., word count, tone, grammar, affective, cognitive, social, punctuation marks, etc.). Results: Significant differences between suicide decedents and living controls were found. Adolescent suicide decedents used more words per sentence, more references to sadness, male individuals, drives, and leisure and fewer verbs and references to they, affiliation, achievement, and power. Limitations: Methodological limitations include the use of only public accounts, small sample size, occasional short posts, and lack of adjustment for multiple testing. Conclusion: Although the sample size is relatively small and only included youth with public accounts, we identified differences in Instagram caption language between adolescents who died by suicide as compared to living controls.

Visuomotor Figure Construction and Visual Figure Delayed Recall and Recognition in Primary Progressive Aphasia

BACKGROUND

Individuals with primary progressive aphasia (PPA) develop visuospatial deficits over time, and those with logopenic variant (lvPPA) are at greatest risk of developing such deficits. However, not all previous studies of visuospatial deficits in PPA have ensured equivalent duration of disease across variants and few have measured deficits longitudinally.

AIMS

The aims of our study were to: 1) investigate differences in baseline visuomotor figure construction, visual figure delayed recall, and figure recognition in PPA variants with similar symptom duration at baseline, and 2) explore patterns of decline in these areas.

METHODS & PROCEDURES

Ninety-three individuals with PPA [39 lvPPA, 24 nonfluent agrammatic PPA (nfaPPA), and 30 semantic variant PPA (svPPA)] were administered the Benson Complex Figure Copy, Benson Complex Figure Delay (Recall), and Benson Figure Recognition. Thirty individuals completed this testing 3 to 47 months post baseline.

OUTCOME & RESULTS

Participants with lvPPA and svPPA showed lower mean scores than those with nfaPPA on visual figure delayed recall at baseline, even though there were no differences in estimated time from disease onset or correlation with disease severity as reflected by naming performance, F(2, 90) = 5.78, p < .004. Those with nfaPPA performed significantly better than those with lvPPA, Tukey HSD p < .05, and those with svPPA, Tukey HSD p < .01. There were no differences between variants in rate of decline in visuomotor figure construction, visual figure delayed recall, and figure recognition.

CONCLUSIONS

These findings revealed relatively spared visuospatial memory in nfaPPA, which may aid in the differential diagnosis of PPA and contribute to designing therapy or compensatory strategies.