How our expectations can make us hallucinate: the neural mechanisms underlying perception

Project Information

EXPECTPERCEPT

Grant agreement ID: 948548

DOI

10.3030/948548

EC signature date 3 December 2020

Start date 1 August 2021

End date 31 December 2026

Funded under

• EXCELLENT SCIENCE - European Research Council (ERC)

Total cost

€ 1 484 575,00

EU contribution

€ 1 484 575,00

Coordinated by

UNIVERSITY COLLEGE LONDON
United Kingdom

Periodic Reporting for period 2 - EXPECTPERCEPT (How our expectations can make us hallucinate: the neural mechanisms underlying perception)

Reporting period: 2023-02-01 to 2024-07-31

Summary of the context and overall objectives of the project

The way we perceive the world is strongly influenced by our expectations about what we are likely to see at any given moment. In certain situations – namely when sensory signals are very weak or noisy and expectations are very strong – expectations can even induce hallucinations: seeing an expected stimulus despite its absence. However, the neural mechanisms by which the brain integrates sensory inputs and expectations, and thereby generates the contents of perception, have yet to be established. Previous work, including my own, has demonstrated that processing in the visual cortex is strongly modulated by prior expectations. However, perhaps surprisingly, previous studies have not yet explained how these modulations relate to subjective perception, leading to a lacuna in our knowledge of how the brain supports perception.
The current project aims to study the neural mechanisms underlying subjective perception by using strong visual expectations to induce hallucinations in healthy human participants. It aims to test the hypothesis that, upon presentation of a predictive cue (e.g. a siren), memory systems pre-activate templates of expected stimuli (an ambulance) in the deep layers of visual cortex, leading to biased processing of sensory inputs from the very moment they arrive. The project will test this proposal by addressing three complimentary questions: 1) How do expectations filter perception? 2) What is the computational architecture underlying perceptual inference? 3) What is the neural source of expectations? These questions will be addressed by combining psychophysical tasks probing participants’ perception with neuroimaging tools with exquisite spatial (high-field fMRI) and temporal (MEG) resolution. The overarching aim of this research is to provide a mechanistic account of subjective perception. Ultimately, these insights may improve our understanding of clinical disorders characterised by aberrations in perception, such as psychosis.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

In the first 30 months of the project, we have conducted multiple human neuroimaging studies investigating subjective perception as outlined above. We have devised and implemented a novel experimental paradigm aimed at testing effects of expectation on false percepts, i.e. seeing things that are not there. We have tested this paradigm using 7T layer-specific fMRI to distinguish feedback and feedforward signaling, and using magnetoencephalograophy (MEG) to resolve the underlying neural signals with millisecond temporal resolution. These studies revealed neural signals reflecting both expectations and false percepts, but interestingly these two types of signals did not interact. That is, participants had false percepts, but they were not influenced by the expectations we cued. In follow-up large sample online psychophysics work, we found that this is likely because there are two types of expectations at play, expecting to see something vs. nothing, and expecting specific stimulus contents, and that these expectations interact in complex ways.
Additionally, we have used MEG to investigate the brain rhythms involved in signalling expectations. This work revealed that expectations oscillate at alpha rhythms (8-12 Hz), a rhythm that is known to be strongly involved in visual perception. Indeed, this work revealed a direct relationship between the strength of these alpha rhythms and effects of expectation on perceptual performance, suggesting a direct link.
Together, this work has led to fascinating new insights in the neural mechanisms underlying expectation and subjective perception.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

We are currently pursuing the different types of expectation outlined above through computational modelling. Specifically, we are testing whether these results are in line with novel models of perceptual consciousness that distinguish two separate stages in perception inference: deciding whether is something is present, and deciding what that something is.
We are also developing a new MEG study to further pursue the link between alpha rhythms and the effects of expectation on perception. This will bring us closer to understanding the neural circuit underlying visual perception.
Third, we are investigating the neural correlates of a striking auditory speech illusion, where reading a particular word determines what you hear (the ‘brainstorm-green needle’ illusion; https://www.youtube.com/watch?v=bqsLNyQj88Q). Going beyond vision alone and branching into auditory perception is useful for the field, especially since psychosis patients are far more likely to have auditory than visual hallucinations.
Finally, we have started testing a visual illusion wherein perception of current inputs is influenced by subsequent ones, a phenomenon known as postdiction. This is important since perception is often considered as being influenced by previous inputs, but this influence of subsequent inputs points to a profound influence of memory on ongoing perception.