Skip to content

Semantic and Syntactic Computerized Analysis of Free Speech

Semantic and Syntactic Computerized Analysis of Free Speech

Status
Recruiting
Phases
Unknown
Study type
Observational
Source
ClinicalTrials.gov
Registry ID
NCT03525054
Acronym
ASESID
Enrollment
215
Registered
2018-05-15
Start date
2018-05-18
Completion date
2030-05-02
Last updated
2026-06-02

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Diagnosis, Psychiatric, Psychosis, Psychotic Disorders, Schizophrenia and Related Disorders, Schizophrenia Prodromal

Keywords

Psychotic Disorders, Diagnosis, Psychiatric, Schizophrenia Prodromal, Ultra High Risk, Prediction Of Psychosis, Machine Learning, Automated Language Analysis, Semantic Coherence, Syntaxic Complexity

Brief summary

Subtle speech disorganization could be predictive of a transition to schizophrenia of ultra-high-risk patients. The aim of our longitudinal multicenter cohort study is to identify specific linguistic markers of the psychotic transition to validate a french predictive model of this transition using computerized speech analysis techniques

Detailed description

Different scales allow identification of patients at ultra-high-risk to develop psychosis. The current challenge is to identify a predictive marker of transition to schizophrenia. Language disorders, which reflect psyche, could be one of these markers. Computerized speech analysis techniques such as Latent Semantic Analysis (LSA) have already proven their reliability in schizophrenia. These techniques reveal subtle speech disorganization that would be predictive of a clinical transition of ultra-high-risk psychotic patients. A combination of semantic and syntactic analysis could accurately predict the psychotic transition. The aim of our longitudinal multicenter cohort study is to validate this predicitve model in french language as well as identifying specific linguistic markers of the psychotic transition. The initial report including the CAARMS is completed with an audio recording from the initial medical interview. The recording will be transcribed and analyzed by computer following the method of lemmatization and vectorial analysis (LSA). An analysis of the grammatical function (number of words, rate of the various grammatical functions) will also be performed. This first analysis will emerge linguistic markers correlated to transition to psychosis that we will use to construct a predictive model for transition to schizophrenia.

Interventions

None listed

Sponsors

University Hospital, Brest
Lead SponsorOTHER

Study design

Observational model
COHORT
Time perspective
PROSPECTIVE

Eligibility

Sex/Gender
ALL
Age
15 Years to 30 Years
Healthy volunteers
No

Inclusion criteria

* Major and/or minor from 15 to 30 years old * Who alleged a suicidal gesture or idea or behavior that has repercussions in their emotional, social or professional life * If patients receive neuroleptic treatment that impairs cognitive abilities, a one-week wash-out period will be scheduled prior to assessment. * Affiliated with or beneficiary of a health insurance/social security system * Able and willing to provide written informed consent

Exclusion criteria

* History of psychosis * Risk of self-harm or violence not compatible with outpatient treatment * QI\<70 (WAIS) * Neurological disorder or major health problem * Impossibility to interrupt neuroleptic treatment for one week * Refusal to participate

Design outcomes

Primary

MeasureTime frameDescription
Transition to schizophrenia2 yearsDetermined from the CAARMS scale (COMPREHENSIVE ASSESSMENT OF AT RISK MENTAL STATES)

Secondary

MeasureTime frameDescription
Identification of patients at "ultra high risk " for developing schizophreniaDay 0Determined from the CAARMS scale ( COMPREHENSIVE ASSESSMENT OF AT RISK MENTAL STATES)

Countries

France

Contacts

CONTACTChristophe LEMEY, Doctor
christophe.lemey@chu-brest.fr02 98 01 52 06

Outcome results

None listed

Source: ClinicalTrials.gov · Data processed: Jun 3, 2026