Intrinsic monitoring of learning success facilitates memory encoding via the activation of the SN/VTA-Hippocampal loop
- Bellvitge Biomedical Research Institute - IDIBELL, Spain
- University of Barcelona, Spain
- Otto-von-Guericke University, Germany
- Catalan Institution for Research and Advanced Studies, ICREA, Spain
- Otto-von-Guericke-University Magdeburg, Germany
- Center for Behavioral Brain Sciences, Germany
Figures
Figure 1
Schematic overview of trials and conditions in the word-learning paradigm.
(A) Participants completed 10 short encoding blocks. Four pairs of sentences of each condition (M+, M-) and two pairs of non-readable sentences (NR, only for the fMRI experiment) were presented per block. Note that first sentences for each condition are always presented prior to and in a different order than second sentences. (B) Each trial in the fMRI experiment started with a fixation cross lasting 500 ms followed by 6 German words of the sentence for 2 s and 1 s of dark screen. Finally, the new-word was presented for 500 ms. Between trials, there was a variable inter-trial interval of 1 to 6 s (Poisson distribution, Hinrichs et al., 2000). (C) Each trial in Exp. 2 started with a fixation cross lasting 1000 ms, continued with the 7 first Spanish words of the sentence presented for 2 s, and was followed by a 1 s duration dark screen. The new-word was presented for 1000 ms. and was followed by 7 s of a small fixation point presented in the middle of the screen. For first sentences, a new trial was presented after 3 s of dark screen. For second sentences, after this period, a screen with the word Answer appeared and subjects had 3 s to produce a verbal answer. Then, the SAM scales for pleasantness and arousal were sequentially presented (the experiment did not continue until participants provided a rating). Finally, a new second sentence trial started after 3 s of dark screen. M+ (meaning extraction possible during second presentation); M- (correct meaning extraction not possible during second presentation); NR sentences (non readable).
Figure 2
ROI analysis controlling for novelty and task difficulty.
Blue areas depict independent ROIs used for beta-extraction overlaid on a canonical T1-weighted template (VS, HP) or the mean proton density normalized template from all subjects (SN/VTA). Bar graphs show mean beta coefficients within ROI for each condition of interest (M+ correct first sentence, M+ incorrect first sentence, M- correct first sentence, M- incorrect first sentence, M+ correct second sentence, M+ incorrect second sentence, M- correct second sentence, M- incorrect second sentence; the NR condition is shown as a control) with standard error of the mean (dark grey for M+; light grey for M-; white for NR). Paired t-test comparisons for all correct versus incorrect conditions revealed significant differences in all ROIs only for M+ correct versus incorrect trials during the second sentence presentation, when participants successfully learned the meaning of a new-word. L, Left Hemisphere; M+, congruent meaning extraction possible; M-, congruent meaning extraction impossible; NR, non-readable sentences; VS, Ventral Striatum; HP, Hippocampus; SN/VTA, Substantia Nigra/ Ventral Tegmental Area. ***p<0.001; **p<0.005.
Figure 3
ROI analysis of memory effects.
Blue areas depict independent ROIs used for beta-extraction overlaid on a canonical T1-weighted template (VS,HP) or the mean proton density normalized template from all subjects (SN/VTA). Bar graphs show mean beta coefficients within ROI for M+ correct trials in which the learned new-word was still remembered during the test performed after the scanning session (remembered) and for M+ correct trials in which the new-word was not properly recognized in the post-scan test (forgotten) with standard error of the mean (dark grey for M+ correct remembered; light grey for M+ correct forgotten; the M+ incorrect condition is shown in white as a control). Paired t-tests showed greater fMRI activity within all ROIs for remembered than for forgotten words, only during the second sentence presentation. M+, congruent meaning extraction possible; VS, Ventral Striatum; HP, Hippocampus; SN/VTA, Substantia Nigra/ Ventral Tegmental Area. ***p<0.005; **p<0.01; *p<0.05.
Figure 4
Changes in subject-specific connectivity due to individual word-learning success.
Blue areas in A, B, C depict the seeds for connectivity-analysis overlaid on a canonical T1-weighted template (VS,HP) or the mean proton density normalized template from all subjects (SN/VTA). Significant correlations between word-learning (learned new-words during the encoding session which were still correctly recognized in the post-scan test) and the change in connectivity in: (A) the left VS (source) and the left HP; (B) the left HP (source) and the left VS; (C) the left SN/VTA (source) and the left HP and the left VS. Results are shown at a p<0.005 uncorrected threshold, with all main peaks within each cluster surviving a p<0.05 FWE corrected threshold within the ROI. The scatter plots display the correlation between the number of learned-words and the mean change in functional connectivity for each particular cluster of interest. Panel (D) shows, on the left, the overlap at the left VS (blue) between the correlations with the source at the left SN/VTA (red) and at the left HP (light blue). On the right, the overlap (yellow) between the correlations with the source at the left SN/VTA and at the left VS (green) is shown. Neurological convention is used in both images with MNI coordinates at the bottom left of each slice. L, Left Hemisphere; VS, Ventral Striatum; HP, Hippocampus; SN/VTA, Substantia Nigra/ Ventral Tegmental Area.
Figure 5
SCR signals and pleasantness/arousal scales during the encoding phase of the second experiment.
(A) Time-course of normalized skin conductance response associated to M+ and M- correct and incorrect conditions during first (left) and second (right) sentence presentation. Solid lines indicate the averaged SCR signal with the corresponding standard error of the mean (red for M+ correct; green for M+ incorrect; light orange for M- correct; light blue for M- incorrect). (B) Mean (averaged using the signal form seconds 4 to 8) normalized SCR for each condition of interest with standard error of the mean (dark grey for M+; light grey for M-). Paired t-test comparisons for all correct versus incorrect conditions showed significant differences only for M+ correct versus incorrect trials during second sentence presentation, when participants successfully learned the meaning of a new-word. (C) Mean pleasantness/arousal ratings (scale between −4 and 4) for each condition of interest with standard error of the mean (dark grey for M+; light grey for M-). Paired t-test comparisons showed that ratings for pleasantness, but not arousal, were greater for M+ correct trials. See Figure 7A for a replication of these effects. M+, congruent meaning extraction possible; M-, congruent meaning extraction impossible. ***p<0.001.
Figure 6
SCR signals and pleasantness/arousal scales in the second experiment for remembered new-words after a 24-hr retention delay.
(A) Time-course of normalized skin conductance response associated to M+ new-words that were remembered or forgotten in the 24 hr test during the first (left) and second (right) sentence presentation. The M+ incorrect conditions are shown as control. Solid lines indicate the averaged SCR signal with the corresponding standard error of the mean (red for M+ remembered; green for M+ forgotten; light blue for M+ incorrect). (B) Bar graphs depict mean normalized SCR (averaged 4 to 8 s post-stimulus) for each condition of interest with standard error of the mean (dark grey for M+ remembered; light grey for M+ forgotten). The M+ incorrect condition is shown in white as a control. Paired t-test comparisons for remembered versus forgotten conditions showed significant differences only for M+ words during second sentence presentation. (C) Mean pleasantness/arousal ratings (scale between −4 and 4) for M+ correct trials in which the new-word was still remembered during the test performed 24-hr after the encoding session (remembered, dark grey) and for those in which the new-word was forgotten (forgotten, light grey). The M+ incorrect condition is shown in white as a control. Paired t-tests revealed significant differences between remembered and forgotten items for pleasantness, but not for arousal. See Figure 7B for a replication of these effects. M+, congruent meaning extraction possible; M-, congruent meaning extraction impossible. ***p<0.005.
Figure 7
Exp. 3.
(A) Mean pleasantness/arousal/confidence ratings (scale between −4 and 4) for each condition of interest with standard error of the mean (dark grey for M+; light grey for M-). M+ incorrect trials have been divided into incorrect-incongruent (subjects say Incongruent instead of providing a meaning), incorrect-other (subjects provide a wrong meaning) and misses (no answer). M- incorrect trials have been divided into incorrect-other (false alarms: subjects provide a meaning for an incongruent new-word) and misses (no answer). (B) Exp.3: Mean pleasantness/arousal/confidence ratings (scale between −4 and 4) for M+ correct trials in which the new-word was still remembered during the test performed 24-hr after the encoding session (remembered, dark grey) and for those in which the new-word was forgotten (forgotten, light grey). (C) Exp.3: Mean pleasantness/arousal/confidence ratings (scale between −4 and 4) for M+ correct trials in which the new-word was still remembered during the test performed one week after the encoding session (remembered, dark grey) and for those in which the new-word was forgotten (forgotten, light grey).M+, congruent meaning extraction possible; M-, congruent meaning extraction impossible.
Figure 8
The SN/VTA-Hippocampal loop.
In the downward arm of the loop, activation starts when new information that needs to be stored in memory ‘arrives’ at the HP via cortical inputs. A signal is then sent to the SN/VTA through the subiculum of the hippocampus, the VS, and the ventral pallidum. Neurons at the SN/VTA are disinhibited by the arriving signal from the HP, which facilitates their dopaminergic firing. In the upward arm of the loop, dopamine is released back into the hippocampus, which in turn enhances memory formation and learning through long term potentiation processes. We suggest that—in the same manner as an extrinsic reward modulates the HP, VS, and SN/VTA, and promotes memory benefits—the activity within the SN/VTA-HP loop can be induced by an intrinsic reward inherent to the process of learning itself and triggered by an internal monitoring of correct performance; and that this intrinsic reward ultimately promotes the storage of new information into long-term memory via dopaminergic modulation of the midbrain. We hypothesize that the prefrontal cortex is fundamental for self-monitoring of correct performance (see Ripollés et al., 2014 for results showing activity within the inferior, middle and superior frontal gyrus while participants were engaged in the same learning task). PFC, prefrontal cortex.
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Intrinsic monitoring of learning success facilitates memory encoding via the activation of the SN/VTA-Hippocampal loop
eLife 5:e17441.
https://doi.org/10.7554/eLife.17441
