The neural dynamics of perceptual priors in audition

Project Information

PriorDynamics

Grant agreement ID: 101032112

DOI

10.3030/101032112

Project closed

EC signature date 19 April 2021

Start date 1 June 2022

End date 31 May 2024

Funded under

• EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions

Total cost

€ 190 734,72

EU contribution

€ 190 734,72

Coordinated by

ECOLE NORMALE SUPERIEURE
France

Periodic Reporting for period 2 - PriorDynamics (The neural dynamics of perceptual priors in audition)

Reporting period: 2023-06-01 to 2024-05-31

Summary of the context and overall objectives of the project

Seeing and hearing the world around us is so effortless that we rarely notice how noisy and ambiguous the information is that our eyes and ears convey. How do we manage to rapidly generate a truthful representation of our sensory environment given such uncertainty? Helmholtz’s study of optical illusions led to the crucial insight that resolving perceptual ambiguities involves prior knowledge, possibly derived from past sensory experience that allows observers to unconsciously infer stimulus properties in a fast and efficient manner. A great amount of research has been devoted to uncovering the mechanisms by which prior information influences perceptual decisions, mostly in the visual modality. However, what is discovered for vision may or may not be a generic mechanism for perception, even though it is often presented as such. Here, we investigate how the auditory system uses information from the recent past to overcome signal noise and ambiguity.

Combining novel behavioural methods with advanced neuroimaging techniques, we aim to determine the neural mechanisms that underlie the propagation and update of prior auditory information in healthy human participants. Specifically, we test the hypothesis that sensory predictions involve neural oscillations at alpha rhythm that mediate the propagation of perceptual priors. The first part of this research project investigates the neural structures underlying such oscillatory mechanism. To this end, we adapt a new time-resolved sampling technique used to examine rhythmic fluctuations in perceptual performance for functional magnetic resonance imaging (fMRI) to explore if cortical and subcortical auditory activations exhibit rhythmic modulations that correlate with oscillations in auditory detection behaviour. The second part of this project examines whether the resolution of perceptual ambiguities by prior contextual information involves similar oscillatory mechanisms. For this purpose, we combine the time-resolved sampling technique with a classic paradigm for inducing ambiguous pitch shifts. The results will shed light on a potentially crucial and yet unknown core process of perception, useful in every interaction we have with the world.

This research project is a joint effort by the École Normale Supérieure in Paris, France, and Chukyo University in Nagoya, Japan.

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

During this first period we established a collaboration with Niigata University in Japan, which guaranteed us access and expertise to their 3 Tesla MRI scanner. In preparation for our planned fMRI study, we worked together with the scientific and technical staff in Niigata to determine the exact scanning protocol in a number of pilot studies. The challenge was how to probe signals related to fast oscillatory activity ~4-12 Hz with a method that has sluggish temporal resolution. We addressed this problem by applying a novel sampling technique originally developed for behavioural measures. This method involves presenting a brief target (e.g. a tone) with random delays relative to a reference event (e.g. a noise burst). The latter serves to realign ongoing brain oscillations to the same phase, making it possible to reveal rhythmic patterns in behaviour.

The next challenge was to adapt this experimental paradigm to the constraints of auditory fMRI that is notoriously affected by scanner noise. Particularly problematic for our study was the fact that scanner noise is rhythmic and, thus, contains frequencies that could interfere with the oscillatory activity we aimed to measure. To overcome this problem, we used a sparse sampling sequence that includes a period of prolonged silence, during which the auditory stimulus is presented, followed by a short interval of functional image acquisition. As a result, sparse sampling yields fewer trials than standard continuous scanning that does not include a silent period. Because lying still in the scanner for a prolonged period of time is very uncomfortable, especially for naive participants, increasing scanning time was not practical.

For that reason, we conducted a separate psychophysics study with the same experimental paradigm but twice the number of trials as in the fMRI. We also took this opportunity to include two other measures that were not planned but may provide important insights into participant perceptual and decisional processes: pupillometry and confidence rating. Recent evidence suggests that changes in pupil size during task performance may reflect fluctuations in arousal and attention which are processes known to modulate brain oscillations. As the mechanisms that control pupillary responses are located within the midbrain, we included subcortical areas in our scanning protocol, so we may be able to link the fMRI results to the pupillometry results. Most importantly, like MRI signals, pupillary responses are slow and therefore present another way to test if fast oscillatory activity can be measured with methods that have low temporal resolution using the sampling technique.

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)

How different brain rhythms contribute to human behaviour is a central question in neuroscience that has been explored extensively with electro- and magnetoencephalogram (M/EEG). Where these rhythms are generated in the brain is a question that is more difficult to answer with M/EEG. Although it is possible to identify the neural sources of brain oscillations by solving the ‘inverse problem’, the results are often unsatisfactory, as the spatial resolution of M/EEG is very poor. Efforts over the past two decades have therefore focused on refining the technique of simultaneously recording EEG and fMRI. Despite recent advances, the EEG data collected in the MRI scanner still suffers from artefacts that cannot be completely removed. Here, we test a new fMRI protocol that does not require EEG and thus could greatly simplify the research on brain rhythms and their neural generators. To further test the feasibility of this protocol, we applied it also to confidence ratings and pupillometry which, to our knowledge, have not been investigated with regard to oscillatory effects to date. Importantly, using the same experimental paradigm as in the fMRI study will allow us to link the confidence and pupillometry results to the fMRI results and thereby gain a better understanding of the neural substrates underlying these behavioural measures, which few investigations are set up to do. Furthermore, to be able to better evaluate participant performance, we included catch trials that comprised of tone targets at either supra- or sub-threshold. The inclusion of these catch trials will also allow us to apply an advanced analysis method, that is, multivariate pattern classification, to decode individual performances to threshold stimuli in auditory activations.

In summary, this research project combines measures of brain activation, pupil modulation and behavioural performance in a way that has not been done before. Therefore, we expect the results to provide new insights into perceptual and decisional mechanisms that underlie the perception of noisy auditory stimuli. There may be no direct socio-economic impact to gain from these results, however, they form an important basis for further research into the neural substrates underlying the rhythmic propagation and update of priors during the perception auditory ambiguous stimuli, which is the focus of the second part of this project.

Experiment design of fMRI study and invidual performance in auditory target detection task