Semantic priming in schizophrenia: systematic review and meta-analysis

Authors:
POMAROL-CLOTET E., et al
Journal article citation:
British Journal of Psychiatry, 192(2), February 2008, pp.92-97.
Publisher:
Cambridge University Press

Increased semantic priming is an influential theory of thought disorder in schizophrenia. However, studies to date have had conflicting findings. Data were pooled from 36 studies comparing patients with schizophrenia and normal controls in semantic priming tasks. Data from 18 studies comparing patients with thought disorder with normal controls, and 13 studies comparing patients with and without thought disorder were also pooled. There was no support for altered semantic priming in schizophrenia as a whole. Increased semantic priming in patients with thought disorder was supported, but this was significant only in comparison with normal controls and not in comparison with patients without thought disorder. Stimulus onset asynchrony (SOA) and general slowing of reaction time moderated the effect size for priming in patients with thought disorder. Meta-analysis provides qualified support for increased semantic priming as a psychological abnormality underlying thought disorder. However, the possibility that the effect is an artefact of general slowing of reaction time in schizophrenia has not been excluded.

Extended abstract:
Author

POMAROL-CLOTET E.; OH T. M. S. S.; LAWS K. R.; et al;

Semantic priming in schizophrenia: systematic review and meta-analysis.

Journal citation/publication details

British Journal of Psychiatry, 192(2), February 2008, pp.92-97.

Summary

People with schizophrenia may experience a pathological heightening of the normal associative patterns between words (semantic priming) that can lead to the incoherent speech known as thought disorder. This meta-analysis of 36 studies finds no support for increased semantic priming in schizophrenia as a whole, but provides qualified support for such a change as a psychological abnormality underlying thought disorder. However, the possibility remains that this finding might derive from the general slowing of reaction times that is characteristic of schizophrenia.

Context

Over 30 studies have now been carried out in this field using tests that generally consist of flashing a string of letters onto a screen (the target), immediately preceded by the flashing of a related or unrelated word (the prime). The time taken to recognise a string of letters as a word is significantly reduced when the prime is related to the target. While some studies have found evidence of increased semantic priming in patients with thought disorder, others have suggested that priming is reduced in schizophrenia as a whole. This meta-analysis aims to clarify matters.

Methods

What sources were used?
PubMed, Medline, PsycINFO and EMBASE (Excerpta Medica) were searched, and the reference lists of relevant primary research and review papers were checked.

What search terms/strategies were used?
The search terms were schizophrenia, priming and semantic priming, and the searches covered the period 1988 to April 2007.

What criteria were used to decide on which studies to include?
Eligible studies reported reaction times in patients with schizophrenia (and normal controls) in any type of semantic priming test. Age and gender matching, and the use of diagnostic criteria for schizophrenia, were not required. Non-English language studies and grey literature such as theses were acceptable.

Who decided on their relevance and quality?
This is not reported. The aim seems to have been to locate all controlled semantic priming studies in populations with schizophrenia.

How many studies were included and where were they from?
Thirty-six studies, including three PhD theses, were included. They are listed in the references to the printed paper, and details are available in an online supplement which gives: patient and control group numbers; mean patient age; mean duration of condition; stimulus onset asynchrony (SOA: the time interval between presentation of the prime and the target); measure of priming; and measure of thought disorder. The online supplement is available via http://bjp.rcpsych.org/content/vol192/issue2/ for those with subscription access to the journal.

How were the study findings combined?
Data from each study were converted into an effect size (Cohen’s d) and the meta-analysis was carried out using DSTAT 1.10, which uses a fixed effects model. Moderator analyses were also carried out on variables such as SOA, age, duration of illness and neuroleptic treatment.

Findings of the review

An analysis of priming in schizophrenia as a whole (i.e. in patients with and without thought disorder) gives a non-significant overall pooled effect size (0.07), suggesting that there is no evidence to support the view that semantic priming is increased in the disorder as a whole.

However, the pooled effect size for the 18 studies comparing patients with thought disorder and normal controls was 0.16, ‘indicating significantly increased priming’. This is a small effect size but when five studies with outlying effect sizes were excluded, it rose to 0.38. Nine studies used an indirect semantic priming test (in which the target word is only indirectly related to the prime) and also produced a small but significant pooled effect size of 0.19, with a still higher result of 0.56 for the six studies comparing thought disorder patients with normal controls.

Moderator analyses of SOA showed that, for the condition as a whole, there was no significant difference between studies with short and long SOAs, and although effect sizes tended to become more negative with increasing SOA, this was also not statistically significant. However, the findings in relation to patients with thought disorder were significant, with a positive pooled effect size for short SOAs and a negative one for long SOAs.

Another potentially moderating effect is the slower than normal reaction times among people with schizophrenia, which is likely to influence the values for semantic priming. A value for the general slowing of reaction time was computed for each study and converted to an effect size. This analysis (of 29 studies) showed that general slowing was a significant moderator of effect size in schizophrenia as a whole, with the greater the slowing the greater the amount of priming. This also held good for the comparison (in 16 studies) between patients with thought disorder and controls.

Among other potentially moderating variables examined, age and duration of illness were not significant, and there were too few studies involving non-medicated patients to make this factor worth examining.

Authors' conclusions

‘Meta-analysis of these studies provides no evidence to support the view that priming (at least direct semantic priming) is increased or decreased in the disorder as a whole, but there is support, albeit qualified, for increased priming in patients with thought disorder. This increase is seen particularly at short SOAs, consistent with an underlying mechanism of increased spread of activation.’ Thus ‘increased semantic priming may be a psychological mechanism underlying thought disorder but is not relevant to the wider clinical picture of the disorder.’

However, the authors warn against too ready an acceptance of the view that thought disorder is associated with increased semantic priming. The effect sizes are generally small, and may be significantly moderated by the general slowing of reaction times characteristic of schizophrenia. Although ‘this does not automatically invalidate the conclusion that semantic priming is increased in patients with thought disorder’, it does require further investigation.

Implications for policy or practice

None are discussed.

Subject terms:
problem solving, psychoses, schizophrenia, communication disorders;
Content type:
systematic review
Link:
Journal home page
ISSN online:
1472-1465
ISSN print:
0007-1250

Key to icons

  • Free resource Free resource
  • Journal article Journal article
  • Book Book
  • Digital media Digital media
  • Journal Journal

Give us your feedback

Social Care Online continues to be developed in response to user feedback.

Contact us with your comments and for any problems using the website.

Sign up/login for more

Register/login to access resource links, advanced search and email alerts