|
|
|
PFC and Episodic Encoding
In the course of a typical day, humans experience
countless events, yet, at the end of the day, only some experiences are
memorable, with many of the day’s events having been forgotten.
Central to understanding memory is determining why some experiences can
be later remembered, whereas others are forgotten. One factor that may
influence memory is the nature and extent of the cognitive control operations
engaged during episodic encoding—that is, the process of transforming
an experience into a durable memory trace such that it can be subsequently
remembered. We have been testing the hypothesis that PFC control mechanisms
guide episodic encoding by bringing task-relevant representations on-line
and elaborating upon them, rendering the accessed and elaborated information
available for input into the MTL memory system.
Our efforts to specify the underpinnings of episodic
encoding have yielded novel insights into the role of cognitive
control in modulating and enhancing episodic learning. Our fMRI
data suggest that distinct regions in PFC impact memory formation
depending on the content of the to-be-learned experience, with
the extent of engagement of PFC processes during an experience
being predictive of whether that experience will be later remembered
or forgotten. For example, our results implicate left VLPFC and
left MTL structures in the encoding of stimuli that can be elaborated
upon at the semantic and phonological levels, and right VLPFC and bilateral
MTL in the encoding of visuo-spatial events [Wagner,
Poldrack et al., 1998; Wagner,
Schacter et al., 1998; Kirchhoff,
Wagner et al., 2000; Davachi,
Maril, & Wagner, 2001 ; Clark & Wagner,
2003; for a review see Wagner,
Koutstaal, & Schacter, 1999; Paller & Wagner,
2002].
We have argued that this content-sensitive pattern of PFC activation reflects
the recruitment of distinct frontal processes that allocate attention to semantic,
phonological, and visuo-spatial representations, guiding their input into declarative
memory [Wagner,
1999; 2002].
Consistent with this hypothesis, we have observed that these content-sensitive
encoding patterns in PFC are accompanied by correlated patterns
in posterior neocortical association areas that putatively represent
the stimulus features being encoded [Kichhoff
et al., 2000; see also,
Dobbins & Wagner,
in press].
An important question surrounding these fMRI-identified
frontal correlates of effective encoding is whether they index
neural computations that play a causal role in learning or are
mere correlates of learning. To begin to test the necessity of
PFC mechanisms for effective encoding, we recently used fMRI-guided
single-pulse TMS to disrupt PFC function during verbal encoding.
Our results revealed that transient disruption of left VLPFC function
during verbal encoding impairs memory performance 20 min later,
whereas disruption of right VLPFC enhanced later memory performance perhaps
by inducing a functional shift to reliance on left-lateralized PFC control
processes that are more effective for verbal encoding [Kahn
et al., 2005]. We are actively pursuing open questions surrounding the role
of PFC control processes in memory formation, including efforts to more
precisely specify the impact of PFC processes in modulating posterior
neocortical and MTL mechanisms.
|
|