NeuroCognitive Imaging Lab at Dalhousie University

What’s Special About Reading Aloud?

This research project is driven by a very simple question: why do we remember things better when we read them aloud, than when we read them silently? There are many possible reasons: when we read aloud, we engage our mouth and larynx (motor system) and we hear ourselves speak (auditory perception). We also simply have to “do” a thing when we read a word aloud that we don’t when we read silently, which may direct more attention to the word. All of these things might strengthen our memory for a word. In our research, we are using functional MRI scanning to examine what is different in the brain when we read aloud. In particular, we are using an advanced analysis technique known as representational similarity analysis (RSA), which allows us to examine the distinct neural “fingerprints” of individual words, in different brain areas.

Below is a more detailed scientific description of this project:

Rationale of the project

The Production Effect is a well-documented phenomenon whereby we are better at remembering words we have spoken aloud, compared to words that we have read silently. At present, this effect has been studied only from a behavioural point of view. Therefore, its neural underpinnings are poorly understood.

Overview of the current study

The overall goal of this project is to better understand the neural underpinnings of the Production Effect and, more generally, the different brain dynamics involved in reading aloud versus reading silently. We use fMRI to record participants’ brain activity while they read words aloud and silently, and later as they are being tested on the words they read. Therefore, we get a record of their brain activity during encoding (i.e., when they are forming memories of the words they are reading) and during retrieval (when they are remembering the words that they read). We are using a few different approaches to analysing our fMRI data. These different approaches can tell us different things about the data. On the one hand, we are using conventional univariate analyses to understand which brain areas are activated when reading aloud and reading silently, and how these differences in activation relate to participants’ memory performance. We are also using multivariate analyses—a relatively new approach to analyzing fMRI data—which can reveal more fine-grained brain dynamics.

Analysis objectives

We are also using multivariate analyses—a relatively new approach to analyzing fMRI data—to understand (1) how brain activity during encoding is re-instated during retrieval, and (2) how different kinds of information are represented in the brain while reading aloud versus reading silently. For example, is articulatory information represented more strongly when reading aloud?

Current state of the study

As of now, we have completed fMRI data collection. The results of our univariate analyses have already been published (Bailey et al., 2021). Our multivariate analyses are ongoing, and we hope to publish further papers reporting on these analyses soon.


Bailey, L.M., Bodner, G.E., Matheson, H.E., Stewart, B.M., Roddick, K., Lambert, A.M., O’Neil, K., Simons, M., Krigolson, O.E., Newman, A.J., & Fawcett, J.M. (2021). Neural correlates of the production effect: An fMRI study, Brain and Cognition.