Latent inhibition, learned irrelevance, and schizotypy: Assessing their relationship

Abstract

Introduction. Previous research suggests that latent inhibition is reduced in patients with acute schizophrenia and in healthy participants with high levels of schizotypic characteristics. Other evidence indicates the disruption of a related effect (learned irrelevance) in patients with acute schizophrenia.

Method. This study used a recently developed latent inhibition procedure, that avoids methodological limitations of previous studies, and a related learned irrelevance procedure to assess the relationship between these phenomena and schizotypic characteristics in undergraduate participants.

Results. Participants preexposed to a letter (S) learnt the predictive relationship between that letter and another letter (X) slower than the relationship between a novel letter and X (a latent inhibition effect). Experiment 1 found reduced latent inhibition in the high schizotypy group after 10, but not 20 preexposures. In Experiment 2, participants preexposed to both S and X learnt a subsequent relationship between them slower than the relationship between a novel letter and X (a learned irrelevance effect). This effect was abolished in participants with high levels of schizotypy.

Conclusions. These results are both the first demonstration of abolished learned irrelevance and of a significant reduction in latent inhibition without employing an explicit masking task in participants with high levels of positive schizotypy.

Keywords:

• Latent inhibition
• Learned irrelevance
• Schizophrenia
• Schizotypy

Acknowledgements

This work was supported by a studentship from the Danish Research Council to Mia Schmidt-Hansen. The authors would like to thank Lisa H. Evans for her advice regarding the design of the experiments reported here and Nicola Gray and Robert Snowden for their comments on this research.

Notes

¹ corresponds in interpretation to R ². An of .01, .06, and .15 constitutes a small, medium, and large effect, respectively (Keppel & Wickens, 2004).

²Subgroups of participants with extremes scores were used rather than regression analyses on the whole sample for the following reason: Mason et al. (1995) report a normative mean for females aged between 16 and 25 years (N=175) for the unusual experiences scale of 11.5 (SD=6.9). In Experiment 1a of the present paper, the whole sample had a mean of 7.09 (SD=5.6), the means of the whole samples in Experiment 1b and 2 were 8.53 (SD=5.96) and 6.44 (SD=4.81), respectively. Under these conditions, regression analyses would be a relatively insensitive tool. Consistent with this suggestion, the regression analyses did not reveal significant relationships between latent inhibition/learned irrelevance and unusual experiences; although the patterns of relationships were in the same direction as revealed by the comparison of the low and high groups.

³When the initial mixed ANOVA with group (no interruption, interruption, walk, and delay) was conducted on the data from only the participants in the low and high groups in Experiment 1a, there was again no effect of group (both Fs < 1), nor did group interact with block or exposure: largest, F(3, 21) = 2.52, MSE=3.87, p=.09, =.27.

⁴When the initial mixed ANOVA with group (no interruption, interruption, walk, and delay) was conducted on the data from only the participants in the low and high groups in Experiment 2, there was again no effect of group (both Fs < 1) and group did not interact with block or exposure (both Fs<1).