The parietal cortex and episodic memory: an attentional account

Abstract

The contribution of the parietal cortex to episodic memory is a fascinating scientific puzzle. On the one hand, parietal lesions do not normally yield severe episodic-memory deficits; on the other hand, parietal activations are seen frequently in functional-neuroimaging studies of episodic memory. A review of these two categories of evidence suggests that the answer to the puzzle requires us to distinguish between the contributions of dorsal and ventral parietal regions and between the influence of top-down and bottom-up attention on memory.

Box 1 Figure

VPC activity shows a U-shaped response as a function of oldness

Figure 1

Subdivisions and connectivity of the posterior parietal cortex. A. The posterior parietal cortex may be divided into the dorsal parietal cortex (DPC), which comprises the lateral cortex including and superior to the intraparietal sulcus and the medial parietal cortex (precuneus—not shown), and the ventral parietal cortex (VPC), which includes the supramarginal gyrus and the angular gyrus. DPC largely correspond to Brodmann Area 7, and VPC, to Areas 40 and 39 B. The parietal cortex has direct anatomical connections with many brain regions, including the frontal lobes (via the superior longitudinal fasciculus—SFL, the fronto-occipital fasciculus—FOF, and the arcuate fasciculus—AC), the temporal lobes (via the middle longitudinal fasciculus—MdlF, and inferior longitudinal fasciculus—ILF). (Fasciculi drawings were adapted from ref.10)

Figure 2

Parietal lesions and episodic memory. A-B. Autobiographical memory was investigated in two patients with bilateral posterior parietal lesions, which were larger in VPC than DPC. A. MRI images showing the lesions outlined in red. B. Participants were impaired when they freely recalled events from their lifetime in as much detail as possible, but not when they answered specific questions about the recalled memories. C-D. Recognition and source memory were compared in patients with left and right VPC lesions. C. The brain slices show the overlap of the lesions. D. Patients were impaired in recollection but not in source memory.

Figure 3

Parietal activation patterns during episodic memory tasks A. Retrieval success activations in the parietal cortex (for the list of studies and coordinate, see supplementary materials online). Event-related fMRI studies that compare hits to correct rejections (CR) typically find activations in posterior parietal regions, including both DPC and VPC. B. Retrieval success activation in left parietal regions tend to become more ventral (smaller Z coordinate) as a function of recognition accuracy (d'). This finding is consistent with the AtoM model. C. Retrieval activation patterns. Retrieval success activity in the parietal cortex is defined as greater activity for hits than for correct rejections (CRs) Recollection-related activity is defined as greater activity for trials associated with recollection than with familiarity (e.g., greater activity for Remember than for Know trials in the Remember/Know paradigm. Perceived oldness is defined as greater activity for or items classified as “old” than for items classified as “new”, regardless of whether these responses are correct or incorrect., . Recollection orienting is defined as greater activity source than for item memory tasks independently of the accuracy of responses (F).-

Figure 4

Ventral-dorsal dissociations in activity. A. In an fMRI study of the Remember/Know paradigm, VPC showed greater activity for Remember than Know trials, whereas DPC showed the opposite pattern. B. In an fMRI study of confidence during recognition memory, VPC showed greater activity for high- than low-confidence hits whereas DPC showed greater activity for low- than high-confidence hits. C. Meta-analysis of parietal activity during episodic retrieval. The figure plots the peaks of activations in two kinds of event-related fMRI studies (for the list of studies and coordinate, see supplementary materials online). A first group of peaks is from studies that identified activity related to recollection or familiarity using the Remember/Know paradigm, distinguishing successful vs. unsuccessful source memory retrieval, or comparing the retrieval of items encoded under deep vs. shallow study tasks. A second group of peaks is from studies that investigated recognition confidence. In general, recollection and high-confidence recognition were associated with VPC activations, whereas familiarity and low-confidence recognition were associated with DPC activations. (Image A was adapted from ref.58, Image B was adapted from ref.61)

Figure 5

Simple graphical description of the Attention to Memory (AtoM) model. Activity in the VPC fluctuates continuously, tracking changes in MTL activity, which in turn reflects the recovery of episodic memories. In contrast, activity in theDPC reflects top-down attentional processes guided by retrieval goals. For example, an earlier segment of the incoming signal may be more relevant to behavioral goals than a later segment. The DPC and VPC interact very closely: the goals maintained by DPC define which targets are relevant, and the targets detected by VPC can alter or change behavioral goals. The attentional processes VPC and DPC contribute to episodic retrieval are the same attentional processes these regions contribute to perception (e.g., dotted arrows in the case of visual attention). (Figure is from ref.72)