JUST PUBLISHED: First-Impression Bias Distorts the Relevance of Sound

Brains are wired to detect the odd one out!

While we are busily engaged in goal-directed activity, our focus of attention is protected from interruption by automatic processes that filter the potential relevance of information around us. The brain’s proficiency as an inferential device is central to this process. By anticipating the likely next-state of the world, the brain minimises resources engaged in processing events that match expectations making us more sensitive to sudden changes that deviate from predictions. For example, you may have noticed this in how readily you tune out consistent noise like the hum of an air conditioner. However, if this noise changes in some way (sounds louder or irregular) your attention may again be drawn to reconsider the relevance and whether action is required.

At the University of Newcastle's Functional Neuroimaging Laboratory, Dr Juanita Todd and colleagues have been studying this process by measuring brain responses to deviations from patterned acoustic input. The brain constantly maps acoustic regularities (even during sleep) to anticipate the most likely properties of sound. Deviations from patterning can capture attention if they elicit sufficiently large responses (e.g., when a deviation is very unlikely). Our studies have centred on understanding the factors that influence deviant response amplitude. Much to our surprise, this relevance-filtering mechanism is subject to a “first-impression” bias. Automatic filters of sound relevance appear to assign potential information value to an unattended sound based on whether it was initially encountered as common and predictable (lower value) or rare an unpredictable (higher value). This first-impression changes the way actual sound probabilities affect deviant response amplitude. We can alter the bias by first asking participants to perform a task with the sounds that they later hear in the unattended sound sequence but curiously we cannot actually abolish the bias as it is remarkably durable.

This research challenges existing models that assume the filter reflects a “low level” system slave to statistical properties of the sound sequences encountered. Instead our research reveals that automated processes distort our internal representation of the environment. We are presently conducting a suite of studies to better understand why and how this bias impacts attention and learning, in particular why first-impressions are given so much weight in our expectations about the world.

For more information about this research, please see the following open-access journal article:

Todd, J., Heathcote, A., Whitson, L. R., Mullens, D., Provost, A., & Winkler, I. (2014). Mismatch negativity (MMN) to pitch change is susceptible to order-dependent bias. Frontiers in Neuroscience, 8 PMID: 25009462