Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 Aug 15;8(1):243.
doi: 10.1038/s41467-017-00180-9.

Sparse Orthogonal Population Representation of Spatial Context in the Retrosplenial Cortex

Affiliations
Free PMC article

Sparse Orthogonal Population Representation of Spatial Context in the Retrosplenial Cortex

Dun Mao et al. Nat Commun. .
Free PMC article

Abstract

Sparse orthogonal coding is a key feature of hippocampal neural activity, which is believed to increase episodic memory capacity and to assist in navigation. Some retrosplenial cortex (RSC) neurons convey distributed spatial and navigational signals, but place-field representations such as observed in the hippocampus have not been reported. Combining cellular Ca2+ imaging in RSC of mice with a head-fixed locomotion assay, we identified a population of RSC neurons, located predominantly in superficial layers, whose ensemble activity closely resembles that of hippocampal CA1 place cells during the same task. Like CA1 place cells, these RSC neurons fire in sequences during movement, and show narrowly tuned firing fields that form a sparse, orthogonal code correlated with location. RSC 'place' cell activity is robust to environmental manipulations, showing partial remapping similar to that observed in CA1. This population code for spatial context may assist the RSC in its role in memory and/or navigation.Neurons in the retrosplenial cortex (RSC) encode spatial and navigational signals. Here the authors use calcium imaging to show that, similar to the hippocampus, RSC neurons also encode place cell-like activity in a sparse orthogonal representation, partially anchored to the allocentric cues on the linear track.

Conflict of interest statement

The authors declare no competing financial interests.

Figures

Fig. 1
Fig. 1
Sparse orthogonal population representation of spatial location in the retrosplenial cortex. a Head-fixed locomotion assay. Mice moved a 150-cm linear treadmill (top) with tactile cues on its surface (bottom). A drop of sucrose water (blue) was delivered at a fixed location for every completed lap. b Lap running behaviour. Movement speed as a function of location for 53 consecutive laps from one experimental session. The animal moved robustly and slowed down or paused most frequently near the reward (as shown by dark colours on the left). c Cellular imaging of neural activity in the retrosplenial cortex (RSC) during head-fixed treadmill running. (top) Illustration of superficial and deep RSC neurons labelled with calcium indicator GCaMP6m (green dots). Calcium imaging was performed with a two-photon microscope through a glass window. (bottom) Tangential view of the labelled superficial RSC neurons with an example imaging field of view (black square). Red lines indicate superior sagittal sinus and transverse sinuses. Scale bar, 1 mm. A: anterior; L: lateral. d Calcium fluorescence signals (top, red) and inferred neural activity (top, black) of six example superficial agranular RSC neurons showing place cell activity; speed and treadmill position are at the bottom. Neural activity was inferred using a fast non-negative deconvolution algorithm 64. e Normalized activity of the six RSC place cells in d as a function of location for multiple laps. The y axis in each colour map corresponds to trial number. Note how neurons were activated as the animal crossed specific locations. Activity was normalized to the time spent at individual locations. f Raster plot showing activation time points for 31 simultaneously imaged RSC place cells, for the same session as in e, together with position (top). Activation time points defined as time points of peak response in each lap for each neuron. Cells ordered by the location that evoked largest responses. Note the repeated sequences of activation during movement and lack of activation when the animal was not moving. g Average normalized activity as a function of location for the 31 RSC place cells shown in f. h Correlation matrix (Pearson correlation coefficient) of population vectors as a function of position for data shown in g. i Correlation matrix (Pearson correlation coefficient) of population vectors as a function of position for data from four mice. (Data from WT mice with AAV1-hSyn-GCaMP6m injections.)
Fig. 2
Fig. 2
Tactile stimuli enhance stability of RSC place cell activity. a (left) Normalized calcium activity of 176 simultaneously imaged RSC place cells on a belt with tactile cues. Cells were ordered by the positions of their peak average activity. Position and speed traces are shown below. Dashed lines and blue drops represent reward delivery. (right) trial-averaged position activity for the 176 RSC place cells shown on the left. Belt diagram (top) and speed traces as a function of position (bottom) are shown. Grey lines, speed traces for individual trials; black line, average speed trace. b Same, for 135 neurons imaged during movement on a belt devoid of salient tactile cues. Note the increased positional jitter of RSC place-cell activity in absence of salient tactile stimuli. c Correlation matrices (Pearson correlation coefficient) of population vectors as a function of position for RSC place cells on the cue belt (left) and on the blank belt (right). d Cumulative distributions of spatial information for all RSC place cells on the cue belt (red) and on the blank belt (blue). (Data from Thy1 GP4.3 transgenic mice.)
Fig. 3
Fig. 3
Similar spatial response properties of RSC and CA1 place cells. a Place-field fraction as a function of place field location on the track. (top) RSC place cells (n = 297); bottom: CA1 place cells (n = 611, electrophysiology and imaging). Error bars: s.e.m. b Cumulative probability distributions of place-field widths for RSC place cells (black, n = 297) and CA1 place cells (red, n = 452, imaging). c Distribution of place field count per cell for RSC (black bars, n = 297) and CA1 (red bars, n = 611, electrophysiology and imaging) place cells. d Population vector correlation (Pearson correlation coefficient) as a function of distance for RSC (black) and CA1 (red) place cells. Shaded areas represent s.d. Note that the periodicity occurred because of the periodicity of the track. (Data from WT mice with AAV1-hSyn-GCaMP6m injections.)
Fig. 4
Fig. 4
RSC place cells are more prevalent in superficial layers. a Diagram of three RSC sub-regions. Sup. agr.: superficial agranular; Deep: deep agranular and granular; Sup. gr.: superficial granular. Green dots represent GCaMP6m labelled neurons. Imaging at different depths revealed neurons in different sub-regions. b Mean population vector correlation as a function of distance for the three sub-regions. Shaded areas represent s.d. c Cumulative probability distributions of place-field widths for superficial agranular (black, n = 123), deep (purple, n = 60), and superficial granular (blue, n = 30) RSC place cells. d Distribution of place field count per cell for superficial agranular (black bars), deep (purple bars) and superficial granular (blue bars) RSC place cells. e Place cell fraction in superficial agranular (black bar), deep (purple bar), and superficial granular (blue bar) RSC. Error bars: s.e.m. (Data from WT mice with AAV1-hSyn-GCaMP6m injections.)
Fig. 5
Fig. 5
Correlated RSC population representations of position upon environmental illumination change and reward relocation. a Spatial tuning curves of three example RSC place cells showing preserved place fields under light (black) and dark (red) conditions. b Population vector correlation (Pearson correlation coefficient) matrix for all RSC place cells under the light and dark conditions. Blue drops indicate reward locations. c (top) Diagram of the reward shift experiment. (bottom) Spatial tuning curves of two example RSC place cells showing preserved place fields when the reward site was shifted. Black lines, spatial tuning curves before reward shift; red lines, spatial tuning curves after reward shift. Blue drops indicate reward locations. d Population vector correlation (Pearson correlation coefficient) matrix for all RSC place cells under original and reward shifted conditions. Blue drops indicate reward locations. (Data from Thy1 GP4.3 transgenic mice a, b and WT mice with AAV1-hSyn-GCaMP6m injections c, d.)

Similar articles

See all similar articles

Cited by 29 articles

See all "Cited by" articles

References

    1. O'Keefe J, Dostrovsky J. The hippocampus as a spatial map. Preliminary evidence from unit activity in the freely-moving rat. Brain Res. 1971;34:171–175. doi: 10.1016/0006-8993(71)90358-1. - DOI - PubMed
    1. McNaughton B, et al. Deciphering the hippocampal polyglot: the hippocampus as a path integration system. J. Exp. Biol. 1996;199:173–185. - PubMed
    1. Burgess N, Recce M, O’Keefe J. A model of hippocampal function. Neural Netw. 1994;7:1065–1081. doi: 10.1016/S0893-6080(05)80159-5. - DOI
    1. Barnes CA, McNaughton BL, Mizumori SJ, Leonard BW, Lin LH. Comparison of spatial and temporal characteristics of neuronal activity in sequential stages of hippocampal processing. Prog. Brain Res. 1990;83:287–300. doi: 10.1016/S0079-6123(08)61257-1. - DOI - PubMed
    1. Wilson MA, McNaughton BL. Reactivation of hippocampal ensemble memories during sleep. Science. 1994;265:676–679. doi: 10.1126/science.8036517. - DOI - PubMed

Publication types

LinkOut - more resources

Feedback