Stimulus content and the neural correlates of source memory

Abstract

It has been suggested that several regions of the brain, including subregions of the medial temporal lobe (MTL) and the posterior parietal cortex, contribute to source memory success in a material-general manner, with most models highlighting the importance of memory process rather than material type. For the MTL in particular, however, increasing evidence suggests that MTL subregions may be specialized for processing different materials, raising the possibility that source memory-related activity may be material-sensitive. Previous fMRI studies have not directly compared source memory activity for different categories of stimuli, and it remains unclear whether source memory effects, in the MTL or elsewhere, are influenced by material. To investigate this issue, young participants were scanned during study while they made semantic judgments about words, pictures of objects and scenes, and during test when they retrieved the context (source) in which these items were studied. Several regions, including the hippocampi, medial and lateral parietal cortex, exhibited source memory effects common to words, objects and scenes, at both study and test. Material-dependent source memory effects were also identified in the left posterior inferior frontal and left perirhinal cortex for words and objects, respectively, at study but not test. These results offer direct support for the hypothesis that the MTL and posterior parietal cortex make material-general contributions to recollection. These results also point to a dissociation between encoding and retrieval with regard to the influence of material on the neural correlates of source memory accuracy, supporting the idea that a relatively small proportion of the activity elicited by a stimulus during encoding is incorporated into an episodic memory representation of the stimulus.

Supplemental Figure 1

Recognition memory effects common across words, objects and scenes as measured at test, displayed on the MNI reference brain. Plots show parameter estimates for the event-related response at the peak maxima of the selected regions for each of the trial types. Error bars depict standard error of the mean difference across participants from left to right: between Source and No Source conditions; No Source and CR conditions; Source and CR conditions [p < 0.0005, uncorrected, with a 5 voxel extent; exclusively masked by Material × Condition interactions at p < 0.05].

Supplemental Figure 2

Recognition memory effects exhibiting differences between stimulus materials as measured at test, shown in selected regions, displayed on the MNI reference brain. Regions identified from recognition memory accuracy (Source/No Source vs. CR) × material interactions. Plots show parameter estimates for the event-related response at the peak maxima of the selected regions for each of the trial types. Error bars depict standard error of the mean difference across participants from left to right: between Source and No Source conditions; No Source and CR conditions; Source and CR conditions [p < 0.0005, uncorrected, with a 5 voxel extent].

Fig. 1

Regions exhibiting differential activity for studied words, objects and scenes, collapsed across subsequent source memory judgment, as measured during study [p < 0.0005, uncorrected, with a 5 voxel extent], displayed on the MNI reference brain.

Fig. 2

Subsequent source memory effects common to words, objects and scenes as measured at study, displayed on the MNI reference brain. Plots show parameter estimates for the event-related response at the peak maxima of the selected regions for trials where source was subsequently recollected with high confidence (Source) and for trials where subsequent source judgments were incorrect or made with low confidence (No Source) for each stimulus category (units arbitrary). Error bars depict standard error of the mean difference across participants between Source and No Source conditions [p < 0.0005, uncorrected, with a 5 voxel extent; exclusively masked by Material × Condition interactions at p < 0.05].

Fig. 3

Source memory effects common to words, objects and scenes as measured at test, displayed on the MNI reference brain. Plots show parameter estimates for the event-related response at the peak maxima of the selected regions for Source, No Source trials and correctly rejected new items (CR) for each stimulus category. Error bars depict standard error of the mean difference across participants from left to right: between Source and No Source conditions; No Source and CR conditions; Source and CR conditions [p < 0.0005, uncorrected, with a 5 voxel extent; exclusively masked by Material × Condition interactions at p < 0.05].

Fig. 4

Subsequent source memory effects exhibiting differences between stimulus materials as measured at study, shown in selected regions, displayed on the MNI reference brain. Plots show parameter estimates for the event-related response at the peak maxima of the selected regions for each of the trial types. Error bars depict standard error of the mean difference across participants between Source and No Source conditions [p < 0.0005, uncorrected, with a 5 voxel extent].