Successful retrieval of competing spatial environments in humans involves hippocampal pattern separation mechanisms

Abstract

The rodent hippocampus represents different spatial environments distinctly via changes in the pattern of "place cell" firing. It remains unclear, though, how spatial remapping in rodents relates more generally to human memory. Here participants retrieved four virtual reality environments with repeating or novel landmarks and configurations during high-resolution functional magnetic resonance imaging (fMRI). Both neural decoding performance and neural pattern similarity measures revealed environment-specific hippocampal neural codes. Conversely, an interfering spatial environment did not elicit neural codes specific to that environment, with neural activity patterns instead resembling those of competing environments, an effect linked to lower retrieval performance. We find that orthogonalized neural patterns accompany successful disambiguation of spatial environments while erroneous reinstatement of competing patterns characterized interference errors. These results provide the first evidence for environment-specific neural codes in the human hippocampus, suggesting that pattern separation/completion mechanisms play an important role in how we successfully retrieve memories.

Keywords: CA1; CA3; MVPA; Memory; Navigation; fMRI; human; neuroscience.

Figure 1.. Experimental design and performance.

(a) Depiction of contextual modifications between environments. Each colored box represents a different target store. Cities 1 & 2 (similar cities) are identical aside from swapped position of stores (purple and teal). City 3 (interference city) shares the same stores as similar cities but in a novel layout. City 4 (distinct city) has a novel layout and stores. (b) During encoding participants completed 4 rounds of navigation and map drawing of each city. (c) Retrieval consisted of 8 blocks of city-specific distance judgments. (d) Retrieval accuracy demonstrates lower performance on city 3. **p<0.01 DOI: http://dx.doi.org/10.7554/eLife.10499.003

Figure 1—figure supplement 1.. Map drawing learning curves.

DOI: http://dx.doi.org/10.7554/eLife.10499.004

Figure 1—figure supplement 2.. City transition map scores.

DOI: http://dx.doi.org/10.7554/eLife.10499.005

Figure 2.. Analysis methods.

(a) Single trial parameter estimates were generated by building a single model with a separate regressor for each trial. (b) Subfields were demarcated manually to create separate ROIs for CA3/DG, CA1, Subiculum, and PHG. (c) The searchlight classifier was trained using single trial estimates from half of the retrieval blocks and tested on the remaining retrieval data. Training/testing was repeated for all searchlight spheres in each subjects MTLs, creating subject specific statistical maps. (d) Within-city similarity was assessed for each ROI by extracting the trial parameter estimates from the subfields and correlating between matched trials of a city’s “A” and “B” retrieval blocks. (e) Between-city similarity was calculated consistent with within-city similarity. DOI: http://dx.doi.org/10.7554/eLife.10499.006

Figure 2—figure supplement 1.. Snapshot of virtual environment.

DOI: http://dx.doi.org/10.7554/eLife.10499.007

Figure 3.. Environment classification.

(a) City classification searchlight revealed a cluster of above chance classification performance throughout much of left CA3/DG and CA1. (b) Pie chart of distribution of voxels in the searchlight showing their predominance in CA3/DG and CA1. (c) Classifier performance of each city revealed above chance performance on cities 1, 2, and 4 and below chance performance on city 3. Further analysis of city 3 classification performance revealed above-chance misclassification of city 3 trials as cities 1 & 2. (d) City 3 (interference city) retrieval performance and city 3 classifier performance were positively correlated. *p<0.05, **p<0.01. DOI: http://dx.doi.org/10.7554/eLife.10499.008

Figure 3—figure supplement 1.. Classifier trained with matched number of trials from each city.

DOI: http://dx.doi.org/10.7554/eLife.10499.009

Figure 3—figure supplement 2.. City 1 & 2 classification results broken down by correctly classified and incorrectly classified as each city.

DOI: http://dx.doi.org/10.7554/eLife.10499.010

Figure 4.. Multivariate pattern similarity analysis (MPS) ofenvironment similarity during retrieval.

(a) Similarity matrix of all pairwise city MPS conditions in CA3/DG. Diagonal depicts within-city and off-diagonal depicts between city MPS conditions. (b) Same as (a) for CA1. (c) Voxel remapping index for CA3/DG (green) and CA1 (blue). Remapping index for each city was the z-transformed contrast between within city and average between cities MPS (see legend below). Left CA3/DG showed overall more remapping than CA1, with significant remapping for Cities 1 & 2 and marginally significant remapping for City 4. Left CA1 showed significant remapping only for City 4. *p<0.05. DOI: http://dx.doi.org/10.7554/eLife.10499.011

Figure 4—figure supplement 1.. Cortical region MPS analysis.

DOI: http://dx.doi.org/10.7554/eLife.10499.012

Figure 5.. Analysis of incorrect and correct interference city trials.

(a) Analysis of interference city trials reveals higher similarity between incorrect city 3 (interfering city) and correct city 1 or 2 trials than between correct city 3 and correct cities 1 and 2 trials in CA3/DG. Control comparisons suggest that this effect could be attributed to interference from cities 1 & 2. Left bar greater than all other bars t(18)>2.2, p<0.04. (b) CA1 did not exhibit similar behavior for incorrect vs correct between-city 3 comparisons. *p<0.05, **p<0.01. DOI: http://dx.doi.org/10.7554/eLife.10499.013

Figure 5—figure supplement 1.. Right hemisphere hippocampal interference city MPS analysis.

DOI: http://dx.doi.org/10.7554/eLife.10499.014

Figure 5—figure supplement 2.. Empirical HRF plotted beside Canonical HRF convolved with 4 s boxcar function (average response time was 3.8 s).

DOI: http://dx.doi.org/10.7554/eLife.10499.015

Author response image 1.. Stable City 1-2 trial classifier results. *p<0.05 Bonferrroni corrected, **p<0.01 Bonferroni corrected.

DOI: http://dx.doi.org/10.7554/eLife.10499.016

Author response image 2.. City 1-2 swapped location trial classifier results. *p<0.05 Bonferrroni corrected, **p<0.01 Bonferroni corrected.

DOI: http://dx.doi.org/10.7554/eLife.10499.017

Author response image 3.. Stable and swapped City 1 and 2 ROI analysis.

DOI: http://dx.doi.org/10.7554/eLife.10499.018