Why does the brain attempt to predict future perceptual input when it can cause such errors? The reason is that in the vast majority of cases expectation improves perception. There is ample evidence from behavioural science that being able to predict the content of an upcoming stimulus improves our ability to successfully perceive it. This is especially true when the signal is degraded in some way, as is often the case with real-world perception (e.g. 1, 2). Despite the advantages of predictive perceptual signaling, the errors caused by expectation are of interest because they can inform us about the limits of our perceptual systems, and how mistakes and dysfunctions within such systems can occur. For example the ‘size-weight illusion’ (3) occurs because we have an expectation that an object’s weight is proportional to it’s size. This expectation causes us to be less accurate at estimating the weight of objects that are unusually high or low in density.
Expectation impacts on our perceptual system by biasing perception towards reporting the expected signal. When expectation is incorrect this can cause us to mistakenly perceive the expected signal when a different but similar signal is present (e.g. 4). For example one might mistake the written word ‘law’ for ‘low’. More startling is the possibility that expectation may cause us to perceive an expected signal when no signal is present. This could explain the occurrence of some hallucinations, where the brain appears to create often complex percepts in the absence of any concordant sensory information. Recently it has been proposed that the auditory hallucinations often experienced by those suffering from psychosis may occur because signals of expected perceptual input become misperceived as genuine sounds (5). Indeed misperceptions due to expectation may be more likely in those with psychosis because the cognitive dysfunctions they suffer make their perceptual predictions less accurate than those of healthy individuals (6)
Previous research has shown that expectation can cause misperceptions when the signal actually present has some similarity with the expected signal (1,4). Further to this it has been shown that erroneous expectation as to the nature of a spoken word can cause a missing phoneme to be heard from within that word (7). This suggests that expectation can cause the perception of speech during periods when no speech at all is present. However hallucinated speech tends to be far more complex then just hearing individual phonemes. Furthermore the (genuine) presence of the rest of the word was required in order to generate the illusionary perception of the missing phoneme in past research. In contrast auditory hallucinations can occur when little or no concordant sensory information related to the hallucinated percept is present. Thus the finding that expectation can generate the perception of missing phonemes doesn’t readily suggest that perceptual expectation could be a cause auditory hallucinations.
Is it possible that the impact of expectation on perception is such that it can cause entire words to be perceived when they are in fact absent? A study recently published in the British Journal of Psychology tested this possibility (8). To mimic the auditory-verbal hallucinations common in psychosis, healthy participants were asked to report whether the final word of a spoken sentence was present when that final word had either been masked or entirely replaced with white noise. A false-positive in this task (reporting speech when only white noise was presented) could be considered analogous to an auditory hallucination in that it represents the perception of auditory-verbal stimulation when no such stimulation is actually present. Crucially during the task the sentence frame (the words preceding the final word) either did or did not generate a specific expectation as to the nature of the final word. This manipulation allowed the effect of expectation on the detection of speech to be tested. Replicating previous findings, the presence of expectation was found to improve the ability to detect speech within the white noise. However it was also found that, despite this improvement, the presence of expectation biased perception to such an extent that it significantly increased the number of false positive responses (8). Thus expectation is capable of generating the perception of spoken words in the absence of any actual speech.
Interestingly in this study the ability of expectation to induce false positive responses was not found to be related to the participant’s level of self-reported hallucination-proneness. This may suggest that while signals of expected input can provide the content for auditory hallucinations, a further deficit must be present in those who hallucinate which allows such signals to erroneously be heard as genuine in naturalistic circumstances. What such deficit(s) might be is the subject of ongoing research which will hopefully lead to a full understanding of the basis of psychotic auditory hallucinations.
1) Krol, M. E., & El-Deredy, W. (2011). When believing is seeing: the role of predictions in shaping visual perception. Q J Exp Psychol (Hove), 64(9), 1743-1771.10.1080/17470218.2011.559587 <Link>
2) Boulenger, V., Hoen, M., Jacquier, C. & Meunier, F. (2011). Interplay between acoustic/phonetic and semantic processes during spoken sentence comprehension: An ERP study. Brain & Language 16, 51-63 <Link>
3) Charpentier, A., 1891. Analyse experimentale de quelques elements de la sensation de poids [Experimental analysis of some elements of weight sensations]. Arch. Physiol. Norm. Pathol. 3, 122–135. <Link>
4) Rahnev, D., Lau, H. & Lange, F. (2011) Prior expectation modulates the interaction between sensory and prefrontal regions in the human brain. The Journal of Neuroscience 31 (29) 10741-10748 <Link>
5) Nazimek, J. M., Hunter, M. D., & Woodruff, P. W. (2012). Auditory hallucinations: expectation-perception model. Med Hypotheses, 78(6), 802-810.S0306-9877(12)00133-8 [pii] 10.1016/j.mehy.2012.03.014 <Link>
6) Fletcher, P. C., & Frith, C. D. (2009). Perceiving is believing: a Bayesian approach to explaining the positive symptoms of schizophrenia. Nature Reviews Neuroscience, 10(1), 48-58. <Link>
7) Samuel, A. G. (1981). Phonemic Restoration – Insights from a New Methodology. Journal of Experimental Psychology-General, 110(4), 474-494 <Link>
8) Hoskin, R., Hunter, M.D., Woodruff, P.W. (2014). The effect of psychological stress and expectation on auditory perception: A signal detection analysis. British Journal of Psychology 105(4) 524-546. <Link>