Episodic Memory Retrieval Functionally Relies on Very Rapid Reactivation of Sensory Information

Abstract

Episodic memory retrieval is assumed to rely on the rapid reactivation of sensory information that was present during encoding, a process termed "ecphory." We investigated the functional relevance of this scarcely understood process in two experiments in human participants. We presented stimuli to the left or right of fixation at encoding, followed by an episodic memory test with centrally presented retrieval cues. This allowed us to track the reactivation of lateralized sensory memory traces during retrieval. Successful episodic retrieval led to a very early (∼100-200 ms) reactivation of lateralized alpha/beta (10-25 Hz) electroencephalographic (EEG) power decreases in the visual cortex contralateral to the visual field at encoding. Applying rhythmic transcranial magnetic stimulation to interfere with early retrieval processing in the visual cortex led to decreased episodic memory performance specifically for items encoded in the visual field contralateral to the site of stimulation. These results demonstrate, for the first time, that episodic memory functionally relies on very rapid reactivation of sensory information.

Significance statement: Remembering personal experiences requires a "mental time travel" to revisit sensory information perceived in the past. This process is typically described as a controlled, relatively slow process. However, by using electroencephalography to measure neural activity with a high time resolution, we show that such episodic retrieval entails a very rapid reactivation of sensory brain areas. Using transcranial magnetic stimulation to alter brain function during retrieval revealed that this early sensory reactivation is causally relevant for conscious remembering. These results give first neural evidence for a functional, preconscious component of episodic remembering. This provides new insight into the nature of human memory and may help in the understanding of psychiatric conditions that involve the automatic intrusion of unwanted memories.

Keywords: EEG; TMS; ecphory; episodic memory; oscillations; retrieval.

Figure 1.

Experimental procedure for Experiments 1 and 2. Pictures of everyday objects were presented to the left or right of fixation at encoding, followed by a response task according to encoding condition (instructed vs noninstructed). Instructed encoding required participants to intentionally encode the presented object and to judge the difficulty to do so. For noninstructed encoding, participants were requested to estimate whether or not the depicted object would fit into a shoebox. During retrieval, all previously shown old items were presented together with the same amount of previously unseen new items. All pictures were shown at the center of the screen to isolate lateralized cortical activity to the reactivation of sensory memory traces established during encoding. An old/new item recognition task was followed by a source memory task, asking for the VF at presentation. The whole procedure was performed twice in each subject: once with instructed, once with noninstructed encoding. In Experiment 1, EEG was measured throughout the experiment. In Experiment 2, rTMS and Sham was applied to the left or right cortical hemisphere during retrieval, switching between these stimulation conditions every 40 trials.

Figure 2.

Posterior electrodes selected for analysis at encoding. Lateral electrodes selected for the initial sliding cluster statistic are depicted in large broken (LH) and dotted (RH) circles. In the topographical cluster analysis comparing LVF and RVF conditions, central electrodes (large solid black) were also included.

Figure 3.

EEG activity in Experiment 1 at sensor level. A, Encoding effects. Top, FDR-corrected results of the sliding cluster statistic, thresholded at p_adj < 0.05, indicating significant power differences between LVF and RVF conditions at respective posterior contralateral sensors (compare Fig. 2) between 200–700 ms and 8–20 Hz (boxed white). Middle, LVF-RVF power differences in the selected time-frequency window. Significant electrode clusters interacting with VF condition represented by black (left hemisphere) and white (right hemisphere) circles. Bottom, Mean power at the left-hemispheric (LH) and right-hemispheric (RH) electrode cluster interaction with VF condition (LVF and RVF). Error bars indicate SEM. *Significant interaction effect (p < 0.05). B, Retrieval effects. Top, Mean difference between contralateral and ipsilateral EEG power for both VF conditions at the LH and RH electrode clusters identified at encoding, showing a contralateral power decrease between 100–200 ms and 10–25 Hz, thresholded at p < 0.05 (Wilcoxon sign rank test). The analysis focused on the time window preceding recollection effects (<500 ms; postrecollection time window masked gray). Middle, LVF-RVF power differences in the selected time-frequency window. Black (left hemisphere) and white (right hemisphere) circles represent electrode clusters identified during encoding. Bottom, Mean power at the left-hemispheric (LH) and right-hemispheric (RH) electrode clusters identified at encoding, interacting with VF condition (LVF and RVF). Error bars indicate SEM. *Significant interaction effect (p < 0.05).

Figure 4.

Hemisphere-specific effects for the LVF-RVF comparison. A, Time-frequency representation of the LVF-RVF difference for left (red circles) and right (blue circles) hemispheric electrode clusters. The 100–200 ms time window selected on the basis of the running Wilcoxon test (Fig. 3B, top) is boxed black (LH) or white (RH). B, Mean difference between LVF and RVF conditions at left and right hemispheric clusters between 100 and 200 ms at the hemisphere-specific peak frequencies (10 and 20 Hz). Only the LH cluster shows a significant difference between VF conditions at 10 Hz, whereas effects for the RH cluster are more pronounced at higher frequencies, peaking at 20 Hz. Error bars indicate SEM. *Significant difference (p < 0.05).

Figure 5.

Cortical sources of EEG power differences between LVF and RVF conditions and between contralateral and ipsilateral hemispheres at encoding (8–20 Hz, 200–700 ms; green) and retrieval (10–25 Hz, 100–200 ms; blue). Interhemispheric differences are backprojected to the cortical hemispheres, reflecting corresponding voxels in each hemisphere. Depicted t values are thresholded at p < 0.01, with maximum values reflecting interhemispheric LVF-RVF differences at p < 0.0005.

Figure 6.

A, Schematic depiction of rTMS in the retrieval trial procedure of Experiment 2 at the left (red) and right (green) maximum cortical source (MNI coordinates: ±40, −78, 0) of 10–25 Hz (100–200 ms) interhemispheric LVF-RVF differences at retrieval as identified in Experiment 1. B, Behavioral results from Experiment 2, showing the difference between TMS-Sham condition effects on source memory performance for items presented in the contralateral and ipsilateral VF during encoding. Error bars indicate SEM. *Significant effect (p < 0.05).