Journal: bioRxiv

Article Title: A hippocampus-accumbens code guides goal-directed appetitive behavior

doi: 10.1101/2023.03.09.531869

Figure Lengend Snippet: (A-D) Training mice on a spatial reward learning task. (A) Schematic of behavioral task. Mice moved on a cue-enriched self-propelled treadmill belt of 360 cm length and obtained a liquid reward when licking at a spout in a hidden reward zone of 30 cm length (yellow). (B) Representative traces of one mouse’s behavior on training day 1 and day 5. (C) Licking in anticipation (Ant.) and reward (Rew.) zone increases over the course of training. (D) Number of rewarded laps per training session ( F (4) = 6.344, GG-correction), anticipatory ( F (4) = 3.803) and reward licking ( F (4) = 4.276, GG-correction) significantly increased over the course of five training days ( n = 18 mice, repeated-measures ANOVA). See also . (E) Schematic of dual-color projection neuron imaging method. Thy1-GCaMP6s mice were injected with AAVrg-Cre in the medial NAc and DIO-mCherry in dHPC. Representative coronal brain slice showing axonal mCherry expression in NAc (AP -1.3; left). Representative coronal brain slice stained with DAPI (blue) of dHPC, showing the outlines of the 3 mm cannula window used for imaging; scale bar represents 1 mm (second left). Field of view of one sample experiment showing Thy1-GCaMP6s expression in green and mCherry expression of putative NAc-projecting neurons in red; outlines show detected components used for analysis (right). (F) Two representative neurons’ raw (“F”), denoised and deconvolved (“C”) and event (“S”) traces; red traces indicate mCherry co-expression (dHPC →NAc ). See also . All data are presented as mean ± SEM. * p < 0.05, ** p < 0.01.

Article Snippet: Then, we used double-floxed inverse open reading frame (DIO) Cre-dependent mCherry under the human synapsin promoter in dHPC (AAV5-hSyn-DIO-mCherry, titer: 1.1 × 10 13 vg/ml, Catalog #50459-AAV5, Addgene, Watertown, USA).

Techniques: Imaging, Injection, Slice Preparation, Expressing, Staining

Journal: bioRxiv

Article Title: A hippocampus-accumbens code guides goal-directed appetitive behavior

doi: 10.1101/2023.03.09.531869

Figure Lengend Snippet: (A) Excerpt from one representative recording session showing behavioral activity (top) and denoised neural activity of identified place cells (bottom) over the course of several minutes from one mouse. White vertical lines mark new laps. Traces are ordered according to whether neurons co-expressed mCherry (red; dHPC →NAc ) or not (green; dHPC − ) and their respective place fields. Black traces are example neurons shown in (B and C). (B and C) Spatial activity of one neuron without (B) and with (C) mCherry co-expression over one session. Calcium events are binned by position on the belt (x axis) and laps (y axis) and show consistent activity at one position on the belt. (D-H) Comparisons of spatial tuning characteristics between dHPC − (green) and dHPC →NAc (red) neurons. NAc-projecting neurons contain a higher proportion of place cells (D, χ 2 (1, 5372) = 6.364); these place cells contain more spatial information per second (E, Welch’s t (186.55) = 4.770), show increased sparsity (F, Welch’s t (218.79) = 3.657), higher reliability (G, probability of firing maximally within their place field per lap; Welch’s t (203.46) = 2.479), and higher in-place field activity (H, Welch’s t (190.32) = 2.884). n = 6 mice, 19 imaging sessions, 5,372 (inc. 444 mCherry-coexpressing) neurons, 1,750 (inc. 169 mCherry-coexpressing) place cells. * p < 0.05, ** p < 0.01, *** p < 0.001. See also .

Article Snippet: Then, we used double-floxed inverse open reading frame (DIO) Cre-dependent mCherry under the human synapsin promoter in dHPC (AAV5-hSyn-DIO-mCherry, titer: 1.1 × 10 13 vg/ml, Catalog #50459-AAV5, Addgene, Watertown, USA).

Techniques: Activity Assay, Expressing, Imaging

Journal: bioRxiv

Article Title: A hippocampus-accumbens code guides goal-directed appetitive behavior

doi: 10.1101/2023.03.09.531869

Figure Lengend Snippet: Three representative fields of view (FOVs) from three different mice are shown, with single-lap spatial calcium activity of each 10 representative neurons, some putatively NAc-projecting. FOV information is shown on top, including numbers of identified dHPC − (green) and dHPC →NAc (red) neurons. FOVs shown are composites of CaImAn maximum local correlation images ( caiman . summary_images . max_correlation_image ) of GCaMP channel (green) and the averaged motioncorrected red channel with NAc-projecting mCherry fluorescence. Contours of ten representative neurons from each FOV are indicated with white outlines and numbers that refer to spatially averaged calcium activity below. Each neuron’s normalized average calcium activity across 45 spatial bins per lap (y axis) is shown with key spatial information values above (SI: spatial information , Sp: Sparsity , Rel: reliability of each neuron’s per-lap maximum activity to occur within the place field). Orange numbers refer to NAc-projecting neurons.

Article Snippet: Then, we used double-floxed inverse open reading frame (DIO) Cre-dependent mCherry under the human synapsin promoter in dHPC (AAV5-hSyn-DIO-mCherry, titer: 1.1 × 10 13 vg/ml, Catalog #50459-AAV5, Addgene, Watertown, USA).

Techniques: Activity Assay, Fluorescence

Journal: bioRxiv

Article Title: A hippocampus-accumbens code guides goal-directed appetitive behavior

doi: 10.1101/2023.03.09.531869

Figure Lengend Snippet: (A) Heat maps of dHPC − (left) and dHPC →NAc (right) place cells’ normalized position-binned average calcium events, ordered by place field location. Texture boundaries are indicated as white dashed lines. Yellow rectangle represents reward zone. (B Example place fields with edges near belt texture boundaries. Primary place fields are indicated by color fill. Triangles mark start (no fill) and end (fill) points. Dashed black lines represent texture boundaries. (C and D) Place field edges accumulate near texture boundary areas. Histograms of dHPC − (green) and dHPC →NAc (red) neurons’ place field start (C) and end (D). Dotted line and shade represent average and 95th CI of 1000x randomly shuffled place fields. Both dHPC − and dHPC →NAc place field start and end positions are significantly overrepresented at the 99.9th percentile (dotted black lines) compared to a randomly shuffled distribution. Start (χ 2 (1, 5134) = 5.735) and end positions (χ 2 (1, 5217) = 4.397) of dHPC →NAc place fields are furthermore significantly overrepresented compared to the dHPC − population. (E and F) Place cells are overrepresented near reward zone in high success trials. (E) Histograms (bars) and kernel density estimations (KDEs; lines) of place field centers for dHPC − (left) and dHPC →NAc (right) neurons, split into high success trials (green/red) and low success trials (gray). Reward zone (Rew.; yellow) and anticipation zone (Ant.; bright yellow) are indicated as rectangles. (F) Proportion of place fields in reward and vicinity zone is significantly higher in high-success trials (colored bars) compared to low-success trials (gray bars) in NAc-projecting neurons (red) but not in dHPC − neurons (green). 2-way ANOVA, F success (1,1) = 54.918, p < 0.001, F projection (1,1) = 0.958, p = 0.338, F interaction (1,1) = 2.969, p = 0.098. Post-hoc Welch’s t -tests with Bonferroni correction: t dHPC– (3.561) = 3.698; t dHPC→NAc (3.479) = 8.671; t low success (13.629) = 0.093, p = 1; t high success (3.952) = 2.075, p = 0.215. Dashed line represents an even distribution of reward and anticipation zone. (G) A linear classifier shows significantly increased decoding accuracy of reward anticipation zone based on dHPC →NAc neural activity compared to that of sample size-matched dHPC − neurons. Wilcoxon’s t -test, W (9) = 5.0, n = 10 imaging sessions. All data are presented as mean ± SEM. * p < 0.05, ** p < 0.01, *** p < 0.001. See also .

Article Snippet: Then, we used double-floxed inverse open reading frame (DIO) Cre-dependent mCherry under the human synapsin promoter in dHPC (AAV5-hSyn-DIO-mCherry, titer: 1.1 × 10 13 vg/ml, Catalog #50459-AAV5, Addgene, Watertown, USA).

Techniques: Activity Assay, Imaging

Journal: bioRxiv

Article Title: A hippocampus-accumbens code guides goal-directed appetitive behavior

doi: 10.1101/2023.03.09.531869

Figure Lengend Snippet: (A) Extent of place fields of dHPC − (left) and dHPC →NAc (right) place cells, sorted by each place cell’s center of mass (COM). Turquoise represents place field, yellow represents COM, black lines represent left and right borders of place fields. Red dashed line represents beginning of reward zone, black dashed lines represent belt texture borders. (B) Width of place fields is unevenly distributed between dHPC − (green; quartiles 35.6 cm / 51.7 cm / 69.0 cm) and dHPC →NAc (red; quartiles 36.5 cm / 58.9 cm / 69.0 cm) place cells (Kolmogorov-Smirnov test, D = 0.116, p = 0.0299). Gray bars represent belt texture zones (60 cm). (C) Place field centers are not biased with respect to texture boundaries (percentile < 95th, permutation test) and the ratio is not different between neuronal populations (χ 2 (1, 5372) = 1.646, p = 0.1995). (D and E) Place field density between reward and anticipation zones compared to the rest of the belt for dHPC − (D) and dHPC →NAc (E) populations. (F) Reward and anticipation zone overrepresentation correlates with behavioral success (percentage of rewarded laps per session) for both dHPC − (green) and dHPC →NAc (red) populations (Pearson correlation; r dHPC– (15) = 0.622, p = 0.0107; . ns: not significant, * p < 0.05.

Article Snippet: Then, we used double-floxed inverse open reading frame (DIO) Cre-dependent mCherry under the human synapsin promoter in dHPC (AAV5-hSyn-DIO-mCherry, titer: 1.1 × 10 13 vg/ml, Catalog #50459-AAV5, Addgene, Watertown, USA).

Techniques:

Journal: bioRxiv

Article Title: A hippocampus-accumbens code guides goal-directed appetitive behavior

doi: 10.1101/2023.03.09.531869

Figure Lengend Snippet: (A-D) Speed-excited dHPC neurons. (A and B) Representative example of one speed-excited neuron. (A) Sample traces of velocity, position and one neuron’s denoised calcium activity. Note the increased calcium activity in times of high velocity, irrespective of position. (B) Linear regression on average calcium events per velocity bin shows a significant positive relationship (slope = 7.43 × 10 −4 , intercept = -2.097 × 10 −3 , r = 0.937, p = 0.0058). (C) Heatmaps of speed-binned normalized calcium activity of all significantly positively speed-modulated dHPC − (top) and dHPC →NAc (bottom) neurons. (D) Proportions of speed-excited neurons are comparable between dHPC − and dHPC →NAc populations (χ 2 (1, 5372) = 2.565, p = 0.109). (E-H) Speed-inhibited dHPC neurons. (E and F) Representative example of one speed-inhibited neuron. (E) Sample traces of velocity, position and one neuron’s denoised calcium activity. Note the increased calcium activity in times of low velocity, irrespective of position. (F) Linear regression on average calcium events per velocity bin shows a significant negative relationship (slope = -1.0437 × 10 −4 , intercept = 2.814 × 10 −3 , r = -0.933, p = 0.0065). (G) Heatmaps of speed-binned normalized calcium activity of all significantly negatively speed-modulated dHPC − (top) and dHPC →NAc (bottom) neurons. (H) Negatively tuned neurons are overrepresented in the NAc-projecting population (χ 2 (1, 5372) = 13.66, p = 0.00022). Speed-modulated neurons were classified as showing a significant linear regression at p < 0.05 after Benjamini/Hochberg FDR correction. All data are presented as mean ± SEM. ns: not significant, *** p < 0.001.

Article Snippet: Then, we used double-floxed inverse open reading frame (DIO) Cre-dependent mCherry under the human synapsin promoter in dHPC (AAV5-hSyn-DIO-mCherry, titer: 1.1 × 10 13 vg/ml, Catalog #50459-AAV5, Addgene, Watertown, USA).

Techniques: Activity Assay

Journal: bioRxiv

Article Title: A hippocampus-accumbens code guides goal-directed appetitive behavior

doi: 10.1101/2023.03.09.531869

Figure Lengend Snippet: (A-C) Consummatory licking leads to a depression in neural activity. (A) Sample experiment showing reward dispensation (purple bars) and consummatory licking onsets (golden vertical lines) and population calcium activity from dHPC − (green) and dHPC →NAc (red) neurons. Note the robust depression of neural activity during times of reward consumption. (B) Event-triggered average traces around reward delivery onset, including licking and speed as well as population average calcium activity of dHPC − (green) and dHPC →NAc (red) neurons. Gray rectangles indicate time windows for comparing calcium activity shown in (C). (C) Calcium activity is differentially modulated by reward onset (two-way mixed ANOVA; F reward_timing (2, 10740) = 19.380, p (GG-corrected) < 0.001, F projection (1, 5370) = 0.798, p = 0.372, F interaction (2, 10740) = 5.559, p = 0.0039). Post-hoc pairwise t -tests with Bonferroni correction were performed for each interaction and the results indicated with asterisks. (D) Lick modulation (average of each trial’s pre-post difference) is significantly increased in dHPC →NAc neurons (Welch’s t (496.11) = 2.379, p = 0.0177). (E) Lick index (average calcium activity difference during licking) is significantly increased in dHPC →NAc neurons (Welch’s t (454.19) = 2.729, p = 0.0066). All data are presented as mean ± SEM. * p < 0.05, ** p < 0.01, *** p < 0.001.

Article Snippet: Then, we used double-floxed inverse open reading frame (DIO) Cre-dependent mCherry under the human synapsin promoter in dHPC (AAV5-hSyn-DIO-mCherry, titer: 1.1 × 10 13 vg/ml, Catalog #50459-AAV5, Addgene, Watertown, USA).

Techniques: Activity Assay

Journal: bioRxiv

Article Title: A hippocampus-accumbens code guides goal-directed appetitive behavior

doi: 10.1101/2023.03.09.531869

Figure Lengend Snippet: (A-C) Appetitive licking is accompanied by increased neural activity in dHPC →NAc neurons but not dHPC − neurons. (A) Representative example showing reward dispensation (purple bar) and appetitive licking onsets (golden vertical lines) and population calcium activity from dHPC − (green) and dHPC →NAc (red) neurons. Note the robust increase of neural activity around lick onsets in the dHPC →NAc population. (B) Event-triggered average traces around appetitive licking onset, including licking and speed as well as population average calcium activity of dHPC − (green) and dHPC →NAc (red) neurons. Gray rectangles indicate time windows for comparing calcium activity shown in (C). (C) Calcium activity is differentially modulated by appetitive licking onset only in dHPC →NAc neurons; two-way mixed ANOVA; F licktiming (1, 5370) = 2.843, p = 0.0918, F projection (1, 5370) = 43.779, p < 0.001 F interaction (1, 5370) = 7.073, p = 0.0079. Post-hoc t -tests with Bonferroni correction: t dHPC– (4927) = 0.871, p = 0.768; t dHPC→NAc (443) = 2.470, p = 0.0277. (D and E) Example imaging session and traces showing lick-excited and lick-inhibited neurons. (D) Field of view showing spatial profiles of dHPC − (green outlines) and dHPC →NAc (red outlines), some of which are classified as lick-excited (violet fill) or lick-inhibited (dark blue fill); neurons #11 and #118 are highlighted. (E) Behavioral traces and calcium activity of sample neurons #11 (lick-inhibited) and #118 (lick-excited). (F and G) Heatmaps (top) and event-triggered calcium activity averages (bottom) of neurons classified as appetitive licking-excited (F) and licking-inhibited (G). (H) Proportion of appetitive licking-excited neurons is higher in NAc-projecting neurons; χ 2 (1, 5372) = 13.018, p = 0.00031. (I) Proportion of appetitive licking-inhibited neurons is not different between populations; χ 2 (1, 5372) = 1.626, p = 0.202. All data are presented as mean ± SEM. ns: not significant, * p < 0.05, *** p < 0.001. See also .

Article Snippet: Then, we used double-floxed inverse open reading frame (DIO) Cre-dependent mCherry under the human synapsin promoter in dHPC (AAV5-hSyn-DIO-mCherry, titer: 1.1 × 10 13 vg/ml, Catalog #50459-AAV5, Addgene, Watertown, USA).

Techniques: Activity Assay, Imaging

Journal: bioRxiv

Article Title: A hippocampus-accumbens code guides goal-directed appetitive behavior

doi: 10.1101/2023.03.09.531869

Figure Lengend Snippet: (A-C) Tracking of orofacial movements via infrared camera recordings. (A) Example still image from near-infrared camera, sampled at 75 Hz. (B) False-color coded motion energy (pixel-by-pixel intensity difference to previous image) overlaid over sample image from (A). Automatically segmented mouth region is indicated by a purple line. (C) Average motion energy in mouth region increases with licking. Golden line shows analog lick spout signal. (D-F) Experimental approach for optogenetic dHPC →NAc stimulation. (D) Injection schematic. (E) Somatic expression of ChR2-EYFP in dorsal pro-subiculum. (F) Axonal expression of ChR2-EYFP in the NAc, where light fibers are placed (tracts indicated by white dotted lines). (G) Representative examples of mouth motion around onset of optogenetic stimulation in an animal expressing ChR2 (top) or EYFP control (bottom). (H) Trial-averaged mouth motion activity around time of optogenetic stimulation in ChR2 (blue) and EYFP (gray)-expressing mice. Mouth motion is significantly increased with optogenetic stimulation in ChR2 animals ( t (3) = 7.485; p = 0.00494) but not EFYP animals ( t (2) = 1.353; p = 0.309). Paired t -tests; n = 4 mice (ChR2), n = 3 mice (EYFP). (J) Trial-averaged relative velocity around time of optogenetic stimulation in ChR2 (blue) and EYFP (gray)-expressing mice. (K) Velocity is significantly decreased with optogenetic stimulation in ChR2 animals ( t (3) = -3.551; p = 0.0381) but not EFYP animals t (2) = -1.263; p = 0.334. Paired t - tests; n = 4 mice (ChR2), n = 3 mice (EYFP). All data are presented as mean ± SEM. ns: not significant, * p < 0.05, ** p < 0.01.

Article Snippet: Then, we used double-floxed inverse open reading frame (DIO) Cre-dependent mCherry under the human synapsin promoter in dHPC (AAV5-hSyn-DIO-mCherry, titer: 1.1 × 10 13 vg/ml, Catalog #50459-AAV5, Addgene, Watertown, USA).

Techniques: Injection, Expressing, Control, Activity Assay

Journal: bioRxiv

Article Title: A hippocampus-accumbens code guides goal-directed appetitive behavior

doi: 10.1101/2023.03.09.531869

Figure Lengend Snippet: (A and B) Conjunctive coding of space and velocity. (A) Venn diagrams of dHPC − (left) and dHPC →NAc (right) place cells and their overlaps with negatively (ocher) and positively (purple) tuned speed cells. Numbers shown refer to absolute numbers of neurons classified. (B) Proportions of dHPC − (green) and dHPC →NAc (red) place cells (fill) and non-place cells (no fill) that are also significantly speed-modulated. A larger proportion of place cells is speed-inhibited compared to non-place cells (dHPC − : χ 2 (1, 4928) = 522.71; dHPC →NAc : χ 2 (1, 444) = 75.02). dHPC →NAc place cells also have a higher proportion of speed-inhibited cells than dHPC − place cells (χ 2 (1, 1750) = 8.977). Speed-excited neurons are overrepresented in non-place cells compared to place cells (dHPC − : χ 2 (1, 4928) = 20.034; dHPC →NAc : χ 2 (1, 444) = 9.966). dHPC →NAc non-place cells also have a higher proportion of speed-excited cells than dHPC − non-place cells (χ 2 (1, 3622) = 6.255). (C and D) Conjunctive coding of space and licking. (C) Event-triggered average calcium traces for dHPC − (green) and dHPC →NAc (red) place cells (left) and non-place cells (right). (D) Proportions of lick-excited neurons are significantly enriched in dHPC − (χ 2 (1, 4928) = 114.515) and dHPC →NAc (χ 2 (1, 444) = 23.248) place cells compared to non-place cells. The proportion of dHPC →NAc lick-excited place cells is also higher than the proportion of dHPC − lick-excited place cells (χ 2 (1, 1750) = 9.442); there is no difference between non-place cells (χ 2 (1, 3622) = 0.488, p = 0.485). There is no difference in the proportions of lick-inhibited place and non-place dHPC − and dHPC →NAc neurons (χ 2 s, all p > 0.05). (E and F) Conjunctive coding of velocity and licking. (E) Histogram of lick-excited (no fill) and lick-inhibited (fill) cells’ velocity correlations (green: dHPC − ; red: dHPC →NAc ). Note the negative velocity correlations of lick-excited cells and positive velocity correlations for lick-inhibited cells. (F) Proportions of lick-excited (left) and lick-inhibited (right) cells among speed-inhibited (no fill), non-speed-modulated (fill) and speed-excited (fill, black stroke) dHPC − (green) and dHPC →NAc (red) cells. Proportions of lick-excited cells are overrepresented in speed-inhibited cells (dHPC − : χ 2 (2, 4928) = 100.484; dHPC →NAc : χ 2 (2, 444) = 27.608). Lick-excited neurons are further enriched in speed-inhibited dHPC →NAc neurons compared to dHPC − neurons (χ 2 (1, 830) = 6.564). Proportions of lick-inhibited cells are overrepresented in speed-excited cells (dHPC − : χ 2 (2, 4928) = 290.832; dHPC →NAc : χ 2 (2, 444) = 18.825). All data are presented as mean ± SEM. ns: not significant, * p < 0.05, ** p < 0.01, *** p < 0.001.

Article Snippet: Then, we used double-floxed inverse open reading frame (DIO) Cre-dependent mCherry under the human synapsin promoter in dHPC (AAV5-hSyn-DIO-mCherry, titer: 1.1 × 10 13 vg/ml, Catalog #50459-AAV5, Addgene, Watertown, USA).

Techniques:

Journal: bioRxiv

Article Title: A hippocampus-accumbens code guides goal-directed appetitive behavior

doi: 10.1101/2023.03.09.531869

Figure Lengend Snippet: (A) Schematic of the generalized linear model (GLM) using position, velocity, and appetitive licking data as predictors for each neuron’s calcium activity (left). Example modelling approach for a triple conjunctive neuron (right). Upper traces show true calcium activity (downsampled and normalized) of test dataset (thin black line) as well as predictions of original model (thick gray) and one example each for each shuffled feature. Note the reductions in prediction quality when features are shuffled. Bottom histograms show R 2 distributions for 100 shuffled models for each feature. Dotted line represents value of 95th percentile, thick black line represents original model’s R 2 . (B) Increased proportions of dHPC →NAc neurons modulated by position (χ 2 (1, 5372) = 93.634), velocity (χ 2 (1, 5372) = 141.86), and licking (χ 2 (1, 5372) = 10.050). (C) Increased proportions of conjunctive coding in dHPC →NAc neurons for position & velocity (χ 2 (1, 5372) = 163.97), position & licking (χ 2 (1, 5372) = 26.029), velocity & licking (χ 2 (1, 5372) = 27.145), and position & velocity & licking (χ 2 (1, 5372) = 34.993). (D) Venn diagrams showing overlap of neurons GLM-classified as modulated by position, velocity and licking in dHPC − (top) and dHPC →NAc (bottom) neurons. Proportions of n -feature coding neurons in dHPC − (left) and dHPC →NAc (right) populations. Non-coding neurons are overrepresented in dHPC − neurons (χ 2 (1, 5372) = 35.382); single-coding neurons are comparably distributed (χ 2 (1, 5372) = 0.0057); dual-coding (χ 2 (1, 5372) = 61.336) and triple-coding (χ 2 (1, 5372) = 30.447) neurons are overrepresented in dHPC →NAc neurons. (F and G) A linear classifier to decode presence of reward zone. (F) Example true presence of reward zone (thin line) and decoder predictions based on non-conjunctive neurons (pink; top) and conjunctive-coding neurons (dark purple; bottom). (G) Conjunctive-coding neurons allow a linear decoder to classify the presence of reward zone more accurately than non-conjunctive coding neurons (Wilcoxon’s W (18) = 14.0). ** p < 0.01, *** p < 0.001. See also Figure S6.

Article Snippet: Then, we used double-floxed inverse open reading frame (DIO) Cre-dependent mCherry under the human synapsin promoter in dHPC (AAV5-hSyn-DIO-mCherry, titer: 1.1 × 10 13 vg/ml, Catalog #50459-AAV5, Addgene, Watertown, USA).

Techniques: Activity Assay

Journal: bioRxiv

Article Title: A hippocampus-accumbens code guides goal-directed appetitive behavior

doi: 10.1101/2023.03.09.531869

Figure Lengend Snippet: (A) GLMs’ explained variance R 2 is significantly lower for dHPC →NAc neurons (Welch’s t (615.15) = 2.147, p = 0.032). (B) Scaled feature importance (standardized GLM coefficients retrieved by H2O’s varimp() method) is higher in dHPC →NAc neurons for position (Welch’s t (2508.09) = 2.162, p = 0.031) and licking (Welch’s t (1879.75) = 9.328, p < 0.001), but not for velocity (Welch’s t (2609.15) = 0.881, p = 0.378). (C) Mean R 2 differences between full and feature-shuffled models are significantly greater in dHPC →NAc neurons for position (Welch’s t (523.23) = 5.984, p < 0.001), velocity (Welch’s t (497.69) = 7.704, p < 0.001), but not for licking (Welch’s t (552.63) = 0.3626, p = 0.717). All data are presented as mean ± SEM. ns: not significant, * p < 0.05, *** p < 0.001.

Article Snippet: Then, we used double-floxed inverse open reading frame (DIO) Cre-dependent mCherry under the human synapsin promoter in dHPC (AAV5-hSyn-DIO-mCherry, titer: 1.1 × 10 13 vg/ml, Catalog #50459-AAV5, Addgene, Watertown, USA).

Techniques: