CoChat
Sign Up
Cited by 265Schizophrenia research and treatmentMental Health Research TopicsMental Health via Writing

The DOI matches its Crossref record; 34 of the 35 references it cites were found in a registry or an index.

Speech Graphs Provide a Quantitative Measure of Thought Disorder in Psychosis

Open access
Mota, Natalia B., Vasconcelos, Nivaldo A. P., Lemos, Nathalia, Pieretti, Ana C., Kinouchi, Osame, Cecchi, Guillermo A., Copelli, Mauro, Ribeiro, Sidarta·2012·PLoS ONE
DOI 10.1371/journal.pone.0034928PMC PMC3322168
Open this paper in CoChat — free accountOpen PDF

Abstract

BACKGROUND: Psychosis has various causes, including mania and schizophrenia. Since the differential diagnosis of psychosis is exclusively based on subjective assessments of oral interviews with patients, an objective quantification of the speech disturbances that characterize mania and schizophrenia is in order. In principle, such quantification could be achieved by the analysis of speech graphs. A graph represents a network with nodes connected by edges; in speech graphs, nodes correspond to words and edges correspond to semantic and grammatical relationships. METHODOLOGY/PRINCIPAL FINDINGS: To quantify speech differences related to psychosis, interviews with schizophrenics, manics and normal subjects were recorded and represented as graphs. Manics scored significantly higher than schizophrenics in ten graph measures. Psychopathological symptoms such as logorrhea, poor speech, and flight of thoughts were grasped by the analysis even when verbosity differences were discounted. Binary classifiers based on speech graph measures sorted schizophrenics from manics with up to 93.8% of sensitivity and 93.7% of specificity. In contrast, sorting based on the scores of two standard psychiatric scales (BPRS and PANSS) reached only 62.5% of sensitivity and specificity. CONCLUSIONS/SIGNIFICANCE: The results demonstrate that alterations of the thought process manifested in the speech of psychotic patients can be objectively measured using graph-theoretical tools, developed to capture specific features of the normal and dysfunctional flow of thought, such as divergence and recurrence. The quantitative analysis of speech graphs is not redundant with standard psychometric scales but rather complementary, as it yields a very accurate sorting of schizophrenics and manics. Overall, the results point to automated psychiatric diagnosis based not on what is said, but on how it is said.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1371/journal.pone.0034928

Cites

34 of 34 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1 Speech graph analysis in schizophrenia, mania and control reports. A) Subjects were asked to report a recent dream. Each report was transcribed and parsed into canonical grammatical elements (words translated from Portuguese, elements separated by slashes). Parts related to dreaming (blue) were sorted from parts related to waking (red), which were considered deviations from the anchor topic. B) Speech graph from the example shown in A), with edges sequentially numbered. The node “I” appears 3 times in the dream sub-graph (“I walked”, “I found”, “I hugged”), and then once in the waking sub-graph (“I woke up”). C) Speech graph examples representative of the schizophrenics (subject MG), manics (subject AB) and controls (subject OR). Graphs plotted using global energy minimum (GEM). The complete database is available as Supporting Information.
Figure 1 Speech graph analysis in schizophrenia, mania and control reports. A) Subjects were asked to report a recent dream. Each report was transcribed and parsed into canonical grammatical elements (words translated from Portuguese, elements separated by slashes). Parts related to dreaming (blue) were sorted from parts related to waking (red), which were considered deviations from the anchor topic. B) Speech graph from the example shown in A), with edges sequentially numbered. The node “I” appears 3 times in the dream sub-graph (“I walked”, “I found”, “I hugged”), and then once in the waking sub-graph (“I woke up”). C) Speech graph examples representative of the schizophrenics (subject MG), manics (subject AB) and controls (subject OR). Graphs plotted using global energy minimum (GEM). The complete database is available as Supporting Information.CC BY · PubMed Central
Figure 2 Local speech graph measures are significantly different for schizophrenics (S), manics (M) and controls (C). Boxplots of speech graph measures. General measures N (P = 0.0028) and E (P = 0.003). Connectivity-related measures LCC (P = 0.005), LSC (P = 0.0078) and ATD (P = 0.007 and P = 0.016). Recurrence-related measures PE (P = 0.0031 and P = 0.0143), L2 (P = 0.0025) and L3 (P = 0.005 and P = 0.0160). Waking-related measures WN (P = 0.0059) and WE (P = 0.0014). Asterisks indicate statistical significance with Bonferroni correction. All the individual raw data and complete statistical results are presented in Tables S3 and S4, respectively.
Figure 2 Local speech graph measures are significantly different for schizophrenics (S), manics (M) and controls (C). Boxplots of speech graph measures. General measures N (P = 0.0028) and E (P = 0.003). Connectivity-related measures LCC (P = 0.005), LSC (P = 0.0078) and ATD (P = 0.007 and P = 0.016). Recurrence-related measures PE (P = 0.0031 and P = 0.0143), L2 (P = 0.0025) and L3 (P = 0.005 and P = 0.0160). Waking-related measures WN (P = 0.0059) and WE (P = 0.0014). Asterisks indicate statistical significance with Bonferroni correction. All the individual raw data and complete statistical results are presented in Tables S3 and S4, respectively.CC BY · PubMed Central
Figure 3 In comparison with schizophrenics, manic patients had significantly more flight of thoughts (higher rate of dream report interruptions to comment on waking events). Ratios of waking nodes/total nodes and waking edges/total edges across groups. For edges, manics scored significantly higher in this ratio than schizophrenics (Kruskal-Wallis P = 0.0397, Wilcoxon Ranksum test P = 0.0146). A non-significant but similar trend was observed for nodes (Kruskal-Wallis P = 0.0597, Wilcoxon Ranksum test P = 0.0258). Asterisk indicates statistical significance with Bonferroni correction.
Figure 3 In comparison with schizophrenics, manic patients had significantly more flight of thoughts (higher rate of dream report interruptions to comment on waking events). Ratios of waking nodes/total nodes and waking edges/total edges across groups. For edges, manics scored significantly higher in this ratio than schizophrenics (Kruskal-Wallis P = 0.0397, Wilcoxon Ranksum test P = 0.0146). A non-significant but similar trend was observed for nodes (Kruskal-Wallis P = 0.0597, Wilcoxon Ranksum test P = 0.0258). Asterisk indicates statistical significance with Bonferroni correction.CC BY · PubMed Central
Figure 4 Global speech graph measures were not significantly different for the different groups. Using the raw data, medians and variances were very similar across groups.
Figure 4 Global speech graph measures were not significantly different for the different groups. Using the raw data, medians and variances were very similar across groups.CC BY · PubMed Central
Figure 5 Schizophrenics produced significantly less words per report than manics. Boxplots of total number of words per report across groups. (Kruskal-Wallis test across groups, P = 0.0067; Wilcoxon Rank Sum test between schizophrenics and manics P = 0.0006; between manics and controls P = 0.2911, and between schizophrenics and controls P = 0.0650). Asterisk indicates statistical significance with Bonferroni correction.
Figure 5 Schizophrenics produced significantly less words per report than manics. Boxplots of total number of words per report across groups. (Kruskal-Wallis test across groups, P = 0.0067; Wilcoxon Rank Sum test between schizophrenics and manics P = 0.0006; between manics and controls P = 0.2911, and between schizophrenics and controls P = 0.0650). Asterisk indicates statistical significance with Bonferroni correction.CC BY · PubMed Central
Figure 6 Schizophrenics and manics are better distinguished by normalized speech graph measures than by psychometric scales BPRS and PANSS. A) Boxplots of normalized data show differences between schizophrenics and manics for N (P = 0.0104), ATD (P = 0.0051), PE (P = 0.0044), and WE (P = 0.0146). B) Boxplots of BPRS and PANSS scores show significant differences between psychotic patients and controls (S>C with P = 0.0003, M>C with P = 0.0012; for PANSS: S>C with P = 0.0003, M>C with P = 0.0006). No significant differences were found between schizophrenics and manics (BPRS S>M with P = 0.1377; PANSS S>M with P = 0.1108). Asterisks indicate statistical significance with Bonferroni correction. All the individual normalized data and complete statistical results are presented in Tables S3 and S4, respectively.
Figure 6 Schizophrenics and manics are better distinguished by normalized speech graph measures than by psychometric scales BPRS and PANSS. A) Boxplots of normalized data show differences between schizophrenics and manics for N (P = 0.0104), ATD (P = 0.0051), PE (P = 0.0044), and WE (P = 0.0146). B) Boxplots of BPRS and PANSS scores show significant differences between psychotic patients and controls (S>C with P = 0.0003, M>C with P = 0.0012; for PANSS: S>C with P = 0.0003, M>C with P = 0.0006). No significant differences were found between schizophrenics and manics (BPRS S>M with P = 0.1377; PANSS S>M with P = 0.1108). Asterisks indicate statistical significance with Bonferroni correction. All the individual normalized data and complete statistical results are presented in Tables S3 and S4, respectively.CC BY · PubMed Central
Figure 7 Normalized global speech graph measures were significantly different between schizophrenics (S) and manics (M). Boxplots of normalized data show significant differences between schizophrenics and manics for Density (P = 0.007), Diameter (P = 0.0138) and ASP (P = 0.0104) with small variance for both groups. In contrast, control subjects yielded a wide span of values between manics and schizophrenics, with high variance.
Figure 7 Normalized global speech graph measures were significantly different between schizophrenics (S) and manics (M). Boxplots of normalized data show significant differences between schizophrenics and manics for Density (P = 0.007), Diameter (P = 0.0138) and ASP (P = 0.0104) with small variance for both groups. In contrast, control subjects yielded a wide span of values between manics and schizophrenics, with high variance.CC BY · PubMed Central
Figure 8 Speech graph measures provide better differential diagnosis of mania and schizophrenia than standard psychometric scales (BPRS and PANSS). Group sorting using graph measures as inputs to the NB classifier was excellent to separate schizophrenics from manics. The measures used as inputs were N, E and ATD for S x M; N, L1 and L2 for S x C; L1, L2 and L3 for M x C. In contrast, group sorting using BPRS and PANSS total scores as inputs for the classifier was successful in separating controls from psychotic patients (either S or M), but sorting between schizophrenics versus manics (S x M) was poor.
Figure 8 Speech graph measures provide better differential diagnosis of mania and schizophrenia than standard psychometric scales (BPRS and PANSS). Group sorting using graph measures as inputs to the NB classifier was excellent to separate schizophrenics from manics. The measures used as inputs were N, E and ATD for S x M; N, L1 and L2 for S x C; L1, L2 and L3 for M x C. In contrast, group sorting using BPRS and PANSS total scores as inputs for the classifier was successful in separating controls from psychotic patients (either S or M), but sorting between schizophrenics versus manics (S x M) was poor.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 35 references

Research works printed but not linked: 1

  1. MH First, 1990, Structured Clinical Interview for DSM-IV Axis I Disorders – Research Version, Patient Edition (SCID-I/P)

Cited by

Sorted by citations

Items 201–265 of 265

Related

See all

Metadata

Authors
Mota, N. B., Vasconcelos, N. A. P., Lemos, N., Pieretti, A. C., Kinouchi, O., Cecchi, G. A., Copelli, M., & Ribeiro, S.
Affiliations
  • Universidade Federal do Rio Grande do Norte (BR)
  • Universidade de São Paulo (BR)
  • IBM Research - Thomas J. Watson Research Center (US)
  • Universidade Federal de Pernambuco (BR)
Corresponding author
Ribeiro, Sidarta
ORCID
Venue
PLoS ONE
Publisher
Public Library of Science (PLoS)
Published
2012-04-09
Volume
7
Issue
4
Pages
e34928
ISSN
1932-6203 (ISSN-L)
DOI
10.1371/journal.pone.0034928
PMC
PMC3322168
OpenAlex
W2091298722
Cited by
265
References
35
Citations by year
2026: 22 · 2025: 18 · 2024: 24 · 2023: 43 · 2022: 34 · 2021: 21 · 2020: 30 · 2019: 15 · 2018: 13 · 2017: 10 · 2016: 12 · 2015: 4 · 2014: 11 · 2013: 6 · 2012: 2
FWCI
5.17
Type
journal article
Language
en
Open access
Licence
Locations
  • PLoS ONE (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Americanae (AECID Library) (publishedVersion, other-oa)
  • LA Referencia (Red Federada de Repositorios Institucionales de Publicaciones Científicas) (publishedVersion, other-oa)
  • Portuguese National Funding Agency for Science, Research and Technology (RCAAP Project by FCT) (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • Figshare (submittedVersion, cc-by)
Funders
  • Governo Brasil
  • Conselho Nacional de Desenvolvimento Científico e Tecnológico (481506/2007-1)
  • Financiadora de Estudos e Projetos (01.06.1092.00)
  • Universidade Federal do Rio Grande do Norte
Topics
  • Health Sciences › Medicine › Psychiatry and Mental health › Schizophrenia research and treatment
  • Social Sciences › Psychology › Experimental and Cognitive Psychology › Mental Health Research Topics
  • Social Sciences › Psychology › Social Psychology › Mental Health via Writing
Keywords
thought disorder, psychosis, schizophrenia, mania, differential diagnosis, graph theory, speech analysis, flight of ideas, automated psychiatric diagnosis, psychiatric symptoms
MeSH terms
Adult, Female, Humans, Male, Middle Aged, Psychiatric Status Rating Scales, Psychotic Disorders, Schizophrenia, Speech, Thinking, Young Adult

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review