Use of explicit memory cues following parietal lobe lesions

Abstract

The putative role of the lateral parietal lobe in episodic memory has recently become a topic of considerable debate, owing primarily to its consistent activation for studied materials during functional magnetic resonance imaging studies of recognition. Here we examined the performance of patients with parietal lobe lesions using an explicit memory cueing task in which probabilistic cues ("Likely Old" or "Likely New"; 75% validity) preceded the majority of verbal recognition memory probes. Without cues, patients and control participants did not differ in accuracy. However, group differences emerged during the "Likely New" cue condition with controls responding more accurately than parietal patients when these cues were valid (preceding new materials) and trending towards less accuracy when these cues were invalid (preceding old materials). Both effects suggest insufficient integration of external cues into memory judgments on the part of the parietal patients whose cued performance largely resembled performance in the complete absence of cues. Comparison of the parietal patients to a patient group with frontal lobe lesions suggested the pattern was specific to parietal and adjacent area lesions. Overall, the data indicate that parietal lobe patients fail to appropriately incorporate external cues of novelty into recognition attributions. This finding supports a role for the lateral parietal lobe in the adaptive biasing of memory judgments through the integration of external cues and internal memory evidence. We outline the importance of such adaptive biasing through consideration of basic signal detection predictions regarding maximum possible accuracy with and without informative environmental cues.

Fig. 1

Diagrams displaying areas of lesion overlap of the patients with parietal and frontal lobe lesions, manually traced on a structural MRI scan of their brain, normalized to MNI space, and displayed on axial slices of a canonical structural image.

Fig. 2

Signal detection model of optimal criterion placement during explicit memory cueing task. The figure demonstrates the optimal criterion placement for an observer with a d′ of 1 under the current cueing manipulation. In the absence of cues the criterion should be placed at the intersection of the distributions. In contrast, the two criteria show the ideal placement under “Likely Old” and “Likely New” cues.

Fig. 3

Accuracy predictions under the signal detection model of adaptive criterion shifts in response to external cues with a 75% validity. Panel (a) shows the accuracy relationships between cued and baseline performance across observes with a range of d′ values, assuming observers place all criteria ideally. Circles illustrate performance under valid cues (both hits and correct rejections) whereas squares illustrate performance under invalid cues (both hits and correct rejections). Panel (b) demonstrates the model’s predictions if observers have moderate noise in their criterion placement across the three cueing conditions (baseline, valid, and invalid). Panel (c) illustrates the predictions for observers who are insensitive to the external cues. Here performance during valid and invalid cueing is similar to baseline, demonstrating insensitivity to the external cues.

Fig. 4

Empirical data linking baseline and cued performance for parietal patients, control participants, and frontal lobe patients during correct detection of old materials-hits. For these materials, “Likely Old” cues are valid whereas “Likely New” cues are invalid. Dotted lines indicate anticipated relation between cued and baseline performance if individuals are insensitive to cues. Solid blue lines demonstrate relationship between baseline and validly cued performance. Solid red lines demonstrate relationship between baseline and invalidly cued performance.

Fig. 5

Empirical data linking baseline and cued performance for parietal patients, control participants, and frontal lobe patients during correct rejection of new materials. For these materials, “Likely Old” cues are invalid whereas “Likely New” cues are valid. Dotted lines indicate anticipated relation between cued and baseline performance if individuals are insensitive to cues. Solid blue lines demonstrate relationship between baseline and validly cued performance. Solid red lines demonstrate relationship between baseline and invalidly cued performance.