Accumulation processes in memory decisions

Every day we need to remember whether we have seen people or items before and
adjust our decisions accordingly. The goal of the proposed research is to
further our understanding of decisions based on remembered information by
following those as they happen in the brain. To be able to study neural
decision dynamics, we use computational models. These models describe decisions
as a protracted process where information is slowly accumulated, and as soon as
the accumulated information crosses a decision threshold, the participant gives
the response corresponding to the decision threshold that is reached (there is
a decision threshold corresponding to each possible answer. In the context of
the tasks we typically use, those would be thresholds for "yes" and "no"
answers). This evidence accumulation process has also been observed in neural
data, during decisions based on perceptual information. But also decisions for
which the information that is needed to make them is not directly available to
the sense (i.e., it has to be retrieved from memory) can be described with
these mathematical models. However, neural correlates of these models have not
yet been described. In the context of decisions based on information in memory,
not only the dynamics at the moment of the decision are of interest, but also
in the stages of the task that lead up to these decisions, which describe how
information is encoded and maintained. Secondary research questions involve
describing how the neural dynamics of those task stages develops. They will
elucidate how memory and decision systems interact dynamically to ensure good
recognition memory performance.

We will study how a memory decision comes into being and how decisions based on
information that is directly available to the senses may or may not differ from
decisions based on information in memory. The neural correlates of perceptual
decisions have been described extensively, but little is known about how the
brain implements decisions based on information in memory.

To answer this question, we will collect concurrent EEG and fMRI measurements
while participants engage in perceptual and memory tasks matched in difficulty
and stimuli. Participants make repeated decisions, such that we can obtain
sufficient amounts of information to average out the noise in the EEG and fMRI
data. For every decision that a participant makes, we can follow the dynamics
with millisecond-resolution in EEG, yet this gives very little information
about the location in the brain where these processes happen. fMRI in contrast
only gives a single datapoint per decision, but has excellent spatial
resolution. We will use the combined EEG/fMRI data (i.e., combining the spatial
resolution of fMRI with the temporal resolution of EEG) to see whether our
neural data agree with predictions of a model of evidence accumulation for
decisions, and to what extent the neural correlates of evidence accumulation
agree between perceptual and memory tasks.

The experiments will not entail more than minimal risk to the participants. The
study is not intended to benefit the subjects directly. Rather, the data
collected during this study can improve our understanding of how decisions made
based on information in memory manifest itself in the brain.