Periodic Reporting for period 1 - PredInCon (The implementation of predictive coding across the auditory hierarchy)
Periodo di rendicontazione: 2023-11-01 al 2025-10-31
In everyday life, we are constantly faced with noisy, incomplete, and ambiguous sensory information. Yet, we are usually able to perceive our environment quickly and accurately. A growing body of research suggests that the brain achieves this by actively generating predictions about what is likely to happen next and combining these expectations with incoming sensory signals. This predictive way of processing information is thought to play a central role in perception, learning, and decision-making.
At the same time, there is increasing recognition that altered predictive processes may contribute to neurological and psychiatric conditions in which perception becomes unreliable, such as hallucinations or sensory distortions. Improving scientific understanding of how predictions and sensory evidence are combined in the healthy brain is therefore important not only for basic neuroscience, but also for long-term progress in mental health research and clinical innovation.
The overall objective of the PredInCon project was to investigate how expectations and sensory information jointly determine what we consciously perceive, using hearing as a model system. Hearing provides an ideal test case because the auditory environment is highly dynamic and often noisy, requiring the brain to continuously anticipate and interpret incoming sounds. The project aimed to clarify how the brain can remain sensitive both to what is expected and to what is surprising, and how this balance shapes conscious experience.
To address these questions, the project developed a new experimental approach in which participants listened for faint sounds embedded in background noise. Crucially, the likelihood of hearing a particular sound was systematically manipulated, allowing to separate the effects of expectation from the strength of the sensory signal itself. This design made it possible to study how prior knowledge and sensory evidence are combined during perception.
By combining behavioural measurements with magnetoencephalographic data, the project set out to identify the brain processes that support this predictive form of perception. Through this approach, PredInCon contributes to European and international research priorities in brain health, cognitive neuroscience, and mental health by improving understanding of fundamental mechanisms of perception. The project’s results are expected to support the development of better theoretical models of brain function and, in the longer term, to inform research on conditions in which predictive processing may be disrupted. In this way, the project helps address the broader need for a deeper, mechanistic understanding of how the brain constructs conscious experience in complex and uncertain environments.
At the same time, there is increasing recognition that altered predictive processes may contribute to neurological and psychiatric conditions in which perception becomes unreliable, such as hallucinations or sensory distortions. Improving scientific understanding of how predictions and sensory evidence are combined in the healthy brain is therefore important not only for basic neuroscience, but also for long-term progress in mental health research and clinical innovation.
The overall objective of the PredInCon project was to investigate how expectations and sensory information jointly determine what we consciously perceive, using hearing as a model system. Hearing provides an ideal test case because the auditory environment is highly dynamic and often noisy, requiring the brain to continuously anticipate and interpret incoming sounds. The project aimed to clarify how the brain can remain sensitive both to what is expected and to what is surprising, and how this balance shapes conscious experience.
To address these questions, the project developed a new experimental approach in which participants listened for faint sounds embedded in background noise. Crucially, the likelihood of hearing a particular sound was systematically manipulated, allowing to separate the effects of expectation from the strength of the sensory signal itself. This design made it possible to study how prior knowledge and sensory evidence are combined during perception.
By combining behavioural measurements with magnetoencephalographic data, the project set out to identify the brain processes that support this predictive form of perception. Through this approach, PredInCon contributes to European and international research priorities in brain health, cognitive neuroscience, and mental health by improving understanding of fundamental mechanisms of perception. The project’s results are expected to support the development of better theoretical models of brain function and, in the longer term, to inform research on conditions in which predictive processing may be disrupted. In this way, the project helps address the broader need for a deeper, mechanistic understanding of how the brain constructs conscious experience in complex and uncertain environments.
The project designed, implemented, and tested a new experimental task to study how expectations and sensory signals are combined during auditory perception. The task required participants to detect faint sounds embedded in background noise, while the probability of hearing specific sounds was systematically varied. This allowed the project to isolate the influence of expectations from the strength of the sensory input itself.
Following extensive piloting and optimisation, the experimental paradigm was implemented in a large-scale brain recording study using magnetoencephalography (MEG). MEG is a non-invasive technique that measures brain activity with very high temporal precision, making it well suited to studying the rapid neural processes involved in perception. A total of 69 participants took part in the study, resulting in a large and high-quality dataset of behavioural responses and brain activity.
The project carried out comprehensive analyses of both behavioural performance and brain signals. Behavioural results showed that participants were more accurate and faster at detecting sounds that were expected compared to sounds that were unexpected. This demonstrates that prior expectations play a measurable role in shaping conscious perception, even when sensory input is weak or noisy.
At the neural level, the project identified brain responses that were specifically enhanced for unexpected sounds, but only when those sounds were consciously detected. This finding provides important evidence that so-called “surprise” or prediction-related brain signals are closely linked to whether a sound reaches conscious awareness, rather than simply reflecting the physical presence of a stimulus.
Together, these achievements provide new experimental evidence on how predictions and sensory evidence jointly contribute to perception. The project delivered a validated experimental paradigm, a large behavioural and brain imaging dataset, and a set of robust results that clarify how the brain balances expectations and incoming information to construct conscious auditory experience.
Following extensive piloting and optimisation, the experimental paradigm was implemented in a large-scale brain recording study using magnetoencephalography (MEG). MEG is a non-invasive technique that measures brain activity with very high temporal precision, making it well suited to studying the rapid neural processes involved in perception. A total of 69 participants took part in the study, resulting in a large and high-quality dataset of behavioural responses and brain activity.
The project carried out comprehensive analyses of both behavioural performance and brain signals. Behavioural results showed that participants were more accurate and faster at detecting sounds that were expected compared to sounds that were unexpected. This demonstrates that prior expectations play a measurable role in shaping conscious perception, even when sensory input is weak or noisy.
At the neural level, the project identified brain responses that were specifically enhanced for unexpected sounds, but only when those sounds were consciously detected. This finding provides important evidence that so-called “surprise” or prediction-related brain signals are closely linked to whether a sound reaches conscious awareness, rather than simply reflecting the physical presence of a stimulus.
Together, these achievements provide new experimental evidence on how predictions and sensory evidence jointly contribute to perception. The project delivered a validated experimental paradigm, a large behavioural and brain imaging dataset, and a set of robust results that clarify how the brain balances expectations and incoming information to construct conscious auditory experience.
The project goes beyond the current state of the art by providing direct experimental evidence on how expectations and surprise-related brain signals jointly shape conscious perception. While predictive processing theories have been widely discussed, empirical tests that link neural signals to the detectability of challenging sounds has so far been lacking. The experimental approach developed in this project makes it possible to establish a solid link between brain activity and participants subjective experience of hearing (or failing to hear) sounds.
A key advance is the demonstration that brain responses to unexpected sounds are selectively enhanced when those sounds are consciously perceived. This finding provides new support for the idea that prediction-related brain signals are not merely automatic responses to sensory input, but are closely tied to perceptual inference and awareness. This helps refine leading theoretical models by showing how surprise and conscious perception are linked at the neural level.
The project also delivers a novel experimental paradigm and a high-quality dataset that can be reused by other researchers. These resources lower barriers for further research by providing a validated method and openly shared data that can be applied to new questions, populations, or sensory modalities. This supports faster progress in the field and encourages independent replication and extension of the findings.
In the longer term, these results have potential relevance for research on neurological and psychiatric conditions in which perception and belief formation are altered, such as hallucinations or sensory distortions. While the project itself is focused on basic research, its findings provide a foundation for future studies aimed at translating predictive processing theories into clinically relevant frameworks.
The availability of open datasets, documented analysis pipelines, and validated experimental tools will facilitate this next phase of research. Together, these advances position the project’s outputs as valuable building blocks for both fundamental neuroscience and longer-term applied research in brain health and mental health.
A key advance is the demonstration that brain responses to unexpected sounds are selectively enhanced when those sounds are consciously perceived. This finding provides new support for the idea that prediction-related brain signals are not merely automatic responses to sensory input, but are closely tied to perceptual inference and awareness. This helps refine leading theoretical models by showing how surprise and conscious perception are linked at the neural level.
The project also delivers a novel experimental paradigm and a high-quality dataset that can be reused by other researchers. These resources lower barriers for further research by providing a validated method and openly shared data that can be applied to new questions, populations, or sensory modalities. This supports faster progress in the field and encourages independent replication and extension of the findings.
In the longer term, these results have potential relevance for research on neurological and psychiatric conditions in which perception and belief formation are altered, such as hallucinations or sensory distortions. While the project itself is focused on basic research, its findings provide a foundation for future studies aimed at translating predictive processing theories into clinically relevant frameworks.
The availability of open datasets, documented analysis pipelines, and validated experimental tools will facilitate this next phase of research. Together, these advances position the project’s outputs as valuable building blocks for both fundamental neuroscience and longer-term applied research in brain health and mental health.