The lateral septum (LS), which is innervated by all CA fields of the hippocampus, contains place cells, and processes spatial information received from the hippocampus (Bezzi, 2005; Olton et al., 1978; Risold and Swanson, 1997; Wirtshafter and Wilson, 2019; Zhou et al., 1999). It has been hypothesized that this spatial information is sent to downstream areas, such as the hypothalamus and ventral tegmental area (VTA) for use in reinforcement and reward seeking (Jiang et al., 2018; Luo et al., 2011; Swanson et al., 1981; Sweeney and Yang, 2015). The LS has also been implicated in contextual reward seeking and the formation of conditioned place preferences (Jiang et al., 2018; Luo et al., 2011). However, little is known about the septal representation of space during reward seeking and what, if any, transformation place information undergoes from its representation in the hippocampus to the lateral septum remains unknown.

The CA fields of the hippocampus contain place cells that show preferential firing when an animal traverses a specific location, the ‘place field’ (O'Keefe and Dostrovsky, 1971; O'Keefe and Nadel, 1978). Extensive work has been done characterizing these fields, including their size (Fenton et al., 2008; Lyttle et al., 2013), skew (Mehta et al., 2000), dependence on experience (Frank et al., 2004; Lee et al., 2004; Mehta et al., 1997; Mehta et al., 2000; Skaggs and McNaughton, 1998), responsiveness to environment and context (Hasselmo and Eichenbaum, 2005; Knierim, 2002; Knierim and Hamilton, 2011), and their distributions around goal locations (Dupret et al., 2010; Hollup et al., 2001; Kobayashi et al., 1997; Kobayashi et al., 2003; Lee et al., 2006). Spatially specific firing has also been well documented and characterized in other brain areas, including in the entorhinal cortex (Fyhn et al., 2004; Hafting et al., 2005; Quirk et al., 1992), where this firing is believed to play a role in path integration (Fyhn et al., 2007; Knierim et al., 2014; Monaco et al., 2011; Moser et al., 2008), and in the lateral septum (Bezzi et al., 2002; Kita et al., 1995; Leutgeb and Mizumori, 2002; Monaco et al., 2019; Takamura et al., 2006; Wirtshafter and Wilson, 2019; Zhou et al., 1999), though much less is known about the potential role of LS place fields in spatial navigation.

Given the anatomical location of the LS between areas involved in spatial navigation such as the hippocampus, and reward/reinforcement related areas such as the VTA (Luo et al., 2011; Risold and Swanson, 1997), and the established role of the LS in reinforcement seeking (Jiang et al., 2018; Mathieu-Kia et al., 1998; McGlinchey and Aston-Jones, 2018; Olds and Milner, 1954; Oshima and Katayama, 2010; Sotomayor et al., 2005), we hypothesized that LS place fields would be preferentially located at reward locations to a greater extent than hippocampal place fields and contain more reward related information.

In the following experiment, we simultaneously recorded cells from the hippocampal CA1 field and the caudodorsal lateral septum during a spatial navigation task. The caudodorsal LS was chosen as it is the area of the LS most heavily innervated by the hippocampus and known to send the most projections to the VTA (Luo et al., 2011; Risold and Swanson, 1997). We characterized lateral septal place field firing and compared it to that seen in CA1. We found that, although not as common as in the hippocampus, the lateral septum contains a large number of place cells with fields that have properties similar to the place cells of the hippocampus, including similar field size and similar number of fields per cell. However, in contrast to HPC place cells, LS cells are slightly less accurately spatially tuned. Additionally, within a firing field, LS cells ramped up their firing in the direction of reward, and place fields in the lateral septum were preferentially biased toward reward locations compared to the hippocampus. We found that HPC and LS cells with reward proximate fields had significantly higher correlated activity than cells with fields located further from reward. We suggest that reward-related spatial information from the hippocampus is preferentially represented in the LS, and we provide three models by which this could occur. We suggest that this relayed information is used downstream to direct the animal to significant or rewarded locations.

In order to characterize the role of the LS in reward-driven spatial navigation, we recorded the activity of 452 caudodorsal LS units and 178 CA1 hippocampus units in six male Long Evans rats (Figure 1A–B). We specifically targeted the caudodorsal LS as it receives the a large innervation from the hippocampus and is the primary source of LS projections to the VTA (Luo et al., 2011; Risold and Swanson, 1997). Units were recorded when the animals performed a spatial working memory task on a double-sided T maze, used previously to characterize spatial activity in multiple regions, including in the hippocampus (Gomperts et al., 2015; Jones and Wilson, 2005a; Jones and Wilson, 2005b; Siegle and Wilson, 2014), prefrontal cortex (Jones and Wilson, 2005a; Jones and Wilson, 2005b) VTA (Gomperts et al., 2015), and lateral septum (Wirtshafter and Wilson, 2019). In this task, animals are forced to a randomly chosen side of one of the Ts (the forced arm), and must run down a center stem and choose the same side of the opposite T (the choice arm) to be rewarded (Figure 1C–D). The structure of this task results in mirrored turns and track traversals which help control for behavioral variability, allowing us to look specifically as the effect of reward approach without confounding variables such as turn approach.

Figure 1

Download asset Open asset

Place fields are less abundant in the LS as compared to the hippocampus CA1

We first sought to determine the prevalence of place fields in the cdLS, as estimates have varied wildly from one study reporting no fields (Tingley and Buzsáki, 2018), to estimates of about a third to half of all LS cells (Takamura et al., 2006; Zhou et al., 1999). In all recorded LS cells and CA1 principle cells, we evaluated the spatial information content in cell firing using an information measure of bits per spike. We defined a cell with place information to be a cell with a bits per spike cutoff of 0.8 bits/spike, as this cutoff has been used previously for determining spatial firing of non-hippocampal cells (Ji and Wilson, 2007; Markus et al., 1994). We did not include cells with a mean spiking rate of less than 0.05 hz, and also eliminated cells during trials where the animal did not cover the entire track at a speed of at least 12 cm/s. We found that 75.2% (124/165) of CA1 cells (Figure 2A) and 33.6% (127/378) of LS cells (Figure 2B) met or exceeded the 0.8bits/spike threshold, with the average bits/spike of a CA1 cell significantly higher than the average bits/spike of an LS cell (HPC mean 1.54+−1.2 bits/spike, LS mean 0.73+−0.7 bits/spike, two-tailed two sample t-test t(541)=9.42, p<0.001) (Figure 2C) (We also computed mutual information, see Figure 2—figure supplement 1). To ensure that the representation of space was different than would be expected from random Poisson firing, we created 454 artificial LS units using Poisson firing and the mean firing rates of the recorded LS units (Figure 2B, inset). The distribution of the bits/spike for the artificial units was highly significantly different than the distribution of bits/spike for actual units (KS test, p<10⁻¹⁵), and only 7.96% of artificial units had bits/spike measurements of greater than 0.8. The average bits/spike for the artificial units was 0.43, compared to an average value of 0.73 for actual units (two-tailed two sample t-test, t(818)=6.72, p<10⁻¹⁰).

Figure 2 with 1 supplement see all

Download asset Open asset

Because bits/spike is sensitive to differences in firing rate, we also calculated bits per second for the CA1 and LS units. We found a mean of 1.34+−1.4bits/sec for CA1 cells, and 0.82+−1.1bits/sec for LS cells (Figure 2D–E). Units in the hippocampus had a significantly greater mean bits/sec than in the LS (two-tailed two sample t-test t(541)=4.57, p<0.001) (Figure 2F).

We next determined how many cells with bits/spike greater than 0.8 had definable place fields. Place fields boundaries have previously been defined in a multitude of ways, including with a flat firing rate threshold (Rich et al., 2014). Because firing rate for LS place fields may not follow the same criteria as hippocampal place fields and may be contingent on the LS’s innate firing properties, we opted to use a threshold derived from the units’ average firing rate. We identified a place field as connected area at least 15 cm long with a peak firing rate of at least two standard deviations above the unit’s mean firing rate, and the boundaries of the place field to be when firing drops below one standard deviation above mean firing rate (see Methods).

Of the 164 HPC units with bits/spike greater than or equal to 0.8, 104 units had at least one place field, with 63 of these cells with one field, 33 with two fields, seven with three fields, and a single cell with four fields (Figure 3A,C). Of the 127 LS cells with bits/spike greater than 0.8, 100 had at least one place field, with 54 with one field, 33 with two fields, six with three fields, five with four fields, and two with five fields (Figure 3B,D). The distributions of field numbers were not significantly different between CA1 and LS cells (two sample Kolmogorov-Smirnov (KS) test, p>0.5). We have previously shown that LS cells are modulated by speed and acceleration (Wirtshafter and Wilson, 2019). To determine the potential contribution of spatially biased speed and acceleration to our place field results, we performed a multiple linear regression of each spatial bin’s firing rate against the bin’s average speed and acceleration. Only two cells with place fields had an r2 of >0.5, with the median r² value of 0.06 (Figure 3—figure supplement 1), so very little spatial firing could be explained by correlations with speed or acceleration.

Figure 3 with 2 supplements see all

Download asset Open asset

In summary, out of 165 recorded CA1 cells that met spike rate and track coverage criteria, 104 (63.0%) met or exceeded a bits per spike cut off of 0.8 and had at least one place field that met the described criteria, while out of 378 recorded LS cells meeting spike rate and track coverage criteria, 100 (26.5%) met the same criteria (see Figure 3—figure supplement 2 for additional place cell examples and representations of firing rates). Place cells in the CA1 are therefore significantly more abundant than in the cdLS (two-tailed two sample t-test t(541)=8.6, p<0.001).

CA1 and LS place fields have comparable average sizes and are shaped by experience

We next determined if there were differences in size between hippocampal and LS place fields. In order to qualify as a place field, we applied a minimum length standard of at least 15 cm. The average length of a CA1 place field was 29.1+−14.7 cm (Figure 4A,C), while the average length of an LS place field was 28.8+−16.6 cm (Figure 4B,D). The average field lengths were not significantly different (two-tailed two sample t-test t(320)=-0.16, p>0.05).

Figure 4

Download asset Open asset

Previous work has demonstrated that hippocampal place fields stabilize and become more tuned to position with experience (Mehta et al., 1997). We compared the time periods over which HPC and LS place fields become stable. For each pass through a place field, we determined how far the center (determined by maximum spiking) of the place field was from the average place field center. Both the hippocampus (Figure 4E) and LS (Figure 4F) had highly accurate place fields starting with the first pass of the place field, though, on average the HPC is slightly more accurate on the first lap as well as across all laps. (On the first lap, HPC has a mean distance of 2.7 cm versus 4.9 cm for the LS, two-tailed two sample t-test t(319)=-3.1, p<0.005. Across all laps, mean distance of 4.70 cm for the HPC versus 5.12 cm for the LS, two-tailed two sample t-test t(6792) = −2.5, p<0.05). The slight but significant decrease in accuracy from the first to later laps in the HPC can be explained by a slight shift of the place field peak toward the direction of travel (Figure 4G). Fields in both the HPC and the LS significantly increase firing in their place field with experience on the track (Figure 4G–H).

CA1 and LS place fields have different skew and location distributions

We also wondered if bias toward reward location could also be seen within a place field. To determine this, looked at place field skew as a function of travel direction, computing skew for a place field using total within-field firing (for skew during individual laps, see Figure 5—figure supplement 1). We found that, when travelling toward a reward site, cells in the hippocampus were skewed positively toward the direction of travel, while, when traveling away from a reward site, cells were skewed negatively away from the direction of travel, and this difference was significant (two-tailed two sample t-test t(207)=-2.1, p<0.05) (Figure 5A,B, see also Figure 5—figure supplement 2–3). Conversely, when traveling toward a reward site, LS place fields were skewed negatively away from the direction of travel, and when traveling away from a reward site, the fields were skewed positively toward the direction of travel, though this difference is not significant (two-tailed two sample t-test t(246)=-1.7, p=0.09) (Figure 5A,C). Although the HPC and LS both contains uni- and bi- directional place fields, there was no significant difference for values of skew based on whether a place cell was uni or bi directional (both two-tailed two sample t-tests, p>0.05, see Figure 5—figure supplement 4). Because skew can be sensitive to changes in speed, we also computed the firing rate asymmetry index (FRAI, see methods) (Mehta et al., 2000) for HPC and LS cells, and found a highly significant relationship between skew and FRAI values (Figure 5—figure supplement 5), showing that skew cannot be explained by the animal’s speed.

Figure 5 with 5 supplements see all

Download asset Open asset

We then examined the effect of reward proximity on skew for place cells that were in the immediate reward proximity (in the rewarded arms, Figure 5D). Comparing the mean skew values in immediate reward proximity (Figure 5E, see also Figure 5—figure supplements 2–3), skew values were not significantly different for hippocampus traveling to and from reward (two sample two sided t-test, t(74)=-0.59 p>0.05), but were significantly different for LS cells to and from reward (two sample two sided t-test, t(108)=-2.10, p<0.05).

We then looked at field location for HPC and LS place fields to determine which arms of the maze were most represented in the hippocampus and LS (Figure 6A). Place field locations were based on maximum spiking (field center) in the place field. In the hippocampus, 49/154 of place fields occurred on the forced sides, 36/154 on the center stem, and 45/154 on the end of the choice sides (Figure 6B,D). (Force and choice points were excluded as they were difficult to assign to an arm, but 24 place fields occurred at the forced and choice point. For data further subdivided by location see Figure 6—figure supplement 1.) In the LS, 37/168 of place fields occurred on the forced sides, 40/168 on the center stem, and 69/168 on the end of the choice sides (22 occurred at choice points) (Figure 6C,E). The distribution of place fields is significantly different between the hippocampus and the lateral septum (Pearson’s chi² test, X² = 6.03, p<0.05). The LS has about 1.4 times, proportionally, significantly more place fields in the choice side than the HPC (41.1% of total fields in the LS, versus 29.2% of total fields in the HPC, two-tailed two sample t-test t(320)=2.23, p<0.05). In the hippocampus, the number of fields in the choice arms versus the forced arms was not significantly different (two-tailed two sample t-test t(364) = 2.34, p>0.05), while, in the LS, there were significantly more fields in the choice arms compared to the forced arms (two-tailed two sample t-test t(334)=-3.8, p<0.001).

Figure 6 with 2 supplements see all

Download asset Open asset

We wondered if place field location depending on direction of travel; for instance, if it was more likely to see a place field by a reward site after the site had been visited. To determine this, we split fields by direction, based on whether the animal was traveling to or from reward (if a field existed in both directions, we analyzed its parameters in both directions). This resulted in a total of 209 hippocampal place fields (115 toward reward and 94 away from reward, with, out of the total, 133 being unidirectional and 76 being bi directional) and 248 LS place fields (138 toward reward, 110 away from reward, with, out of the total, 177 unidirectional and 71 bidirectional. There was no significant difference of numbers of uni- or bi- directional HPC or LS cells, two-tailed two sample t-test, t(461)=1.377, p>0.05), and no difference in the distribution of HPC or LS place cell locations depending on if a cell was uni- or bi- directional both (KS test, p>0.05). There was no significant different between field sizes based on direction within the LS, within the HPC, or across the LS and HPC (all two-tailed two sample t-test p>0.05).

In the LS, the difference seen in the number of fields on the forced side versus choice side was most stark when the animal was traveling away from reward (two-tailed two sample t-test t(218)=3.8 p<0.001), although there was a clear trend while the animal traveled toward reward (two-tailed two sample t-test t(254)=1.82 p<0.1) (Figure 6G). In the hippocampus, there were never significantly more fields on the choice side (Figure 6F). While LS place fields were more concentrated around reward sites, average place field firing rate did not scale with distance to reward in either the LS or the hippocampus (Figure 6—figure supplement 2).

We also computed the probability of finding a place field as a function of distance from reward (Figure 6H–I). In the hippocampus, there was an increase in the probability of a spatial firing field in the last 60 cm of reward approach (Figure 6H). However, the largest peeks in HPC place field probability were around the forced and choice points of the maze, approximately 200–220 cm and 80–60 cm away from reward, respectively. In the LS, the entire forced arms were highly overrepresented, and the probability of a place field also increased upon reward approach.

In order to determine if HPC place cells were preferentially innervating LS cells, we computed spike train cross correlations between HPC and LS place cells with place fields in similar locations (centers less than or equal to 20 cm apart) (Wilson and McNaughton, 1994). (To adjust for spiking variance, we also computed the cross correlation with a shuffled LS spike train, and subtracted the mean of this cross correlation from the computed pairwise cross correlation.) We found that place cell pairs with fields in the choice side of the maze had a significantly higher mean average correlation (over +−100 ms) than place cell pairs with fields in the middle or forced arm (Figure 7A) (forced versus choice: two-tailed two sample t-test t(67)=-2.2, p<0.05, middle versus choice: two-tailed two sample t-test t(54)=-2.3, p<0.05). For pairs in the choice arm, the maximum value of the cross correlation happened at a mean lag of 20 ms with the HPC leading (Figure 7B, for all cross correlations, see Figure 7—figure supplement 1), within approximately the same time course seen for sharp wave ripple propagation from the HPC to the LS (Wirtshafter and Wilson, 2019). It does not appear that the higher cross correlations for cells on the choice side of the track were due to higher firing rates of cells proximal to reward, as there was no significant difference between the mean and maximum firing rates of LS place cells in all three locations (see Figure 7—figure supplement 2). The higher cross correlations were also not the result of greater proximity of field centers between LS-HPC place cell pairs in the choice arm (see Figure 7—figure supplement 3).

Figure 7 with 3 supplements see all

Download asset Open asset

Finally, we sought to determine if correlations between HPC and LS spiking became stronger throughout a single day’s training (Figure 7C). Comparing pairs of cells within the first third, middle third, and last third of training, there were no significant differences within group (e.g. cross correlation of reward proximate cells in the first third versus the last third of training) (all paired double sided t tests, p>0.05). However, across all three periods of training, cross correlations for reward proximate cells were consistently higher than for more reward-distal cells. It therefore appears that activity correlations between HPC and LS reward-proximate cells are either developed early in training before the task is learned and not lost between training sessions, or they are not developed within training sessions.

In this study, we have completed a comprehensive characterization and analysis of LS place cells and their properties using parameters identical to those used to characterize HPC place cells. We found that, while the percentage of place cells is lower in the LS than in CA1, the observed LS place cells are similar to those in the CA1 in terms of field numbers and field size. We additionally determined that LS do not converge on a place field center with as much precision as CA1 cells. Finally, we found that LS place fields are distributed more toward reward locations than place fields in CA1, and that this distribution is even more biased toward rewarded locations if the animal is travelling away from the reward site. We suggest that these characteristics are the result of selective LS innervation by HPC place cells, which then allows reward related information to be sent to areas downstream of the LS.

Previous reports have varied wildly on the prevalence of place cells in the LS, with estimates ranging from none (Tingley and Buzsáki, 2018), to a third (Leutgeb and Mizumori, 2002; Zhou et al., 1999) to almost half or more of all cells (Takamura et al., 2006). We found that 26.5% of LS cells could be characterized as place fields based on both a bits per spike cutoff and spatial firing characteristics. The difference in estimates can be accounted for by several variables: first, using a stringent bits/spike and field size criteria eliminates many cells that may fit one criteria or the other, or that may have very small fields. Additionally, as we found that LS place fields tended to be clustered around reward locations, the presence, absence, and location of reward could have affected the total number of place fields. Finally, the caudodorsal LS receives the densest innervation from the hippocampus (Swanson, 1977), so more ventral areas of the LS may have fewer place fields, a result that has been previously observed (although place fields have been found throughout the entirety of the LS) (Takamura et al., 2006).

LS place field sizes are comparable to HPC field sizes, and both HPC and LS cells increase place field spiking with experience. However, LS place cells are slightly but significantly less accurate than HPC place fields (Figure 4E–H). This may be the result of LS cells incorporating other parameters beyond current spatial location. We have previously shown that LS cells’ firing rate may be affected by an animal’s speed or acceleration (Wirtshafter and Wilson, 2019), which may decrease the precision of place field firing. The incorporation of proximity to reward location, discussed below, may also serve to modulate LS place cell firing beyond the animal’s current location.

We observed that the distribution of place fields in the lateral septum was more biased toward the rewarded locations of the maze than the distribution of place fields in CA1 (Figure 6), and that, unlike HPC place fields, LS place fields tended to skew toward reward direction regardless of the direction of travel, particularly when close to rewarded locations (Figure 5). Previous work has also established that the LS can represent both reward delivery and reward locations (Nerad and McNaughton, 2006; Wirtshafter and Wilson, 2019). There are many possible ways this result could be accounted for at the circuit level: First, reward information may be introduced into the lateral septum from an area outside of the hippocampus. The LS has connections, often reciprocal, with many brain areas that respond to reinforcing stimuli, including the VTA (Gomperts et al., 2015; Jiang et al., 2018; Luo et al., 2011; Roesch et al., 2007), hypothalamus, which is involved with the LS in a circuit that regulates feeding (Sweeney and Yang, 2016), ventral hippocampus (Zhou et al., 2019), and amygdala (Krettek and Price, 1978; Sheehan et al., 2004). It is possible that information about reward location is being sent from, for example, the VTA to the LS, rather than vice versa. However, we think it unlikely that this circuit accounts for the all, or the majority, of reward proximate place firing in the LS for several reasons: First, reward related cells in the VTA are specifically reactivated during hippocampal replay and sharp wave ripples at a time course consistent with the signal first travelling through the LS from the hippocampus (Gomperts et al., 2015; Wirtshafter and Wilson, 2019). Additionally, cells in the LS that change firing during a reward period tend to show signatures of hippocampal interactions, such as theta coherence and increased firing during hippocampal sharp-wave ripples (Wirtshafter and Wilson, 2019). Finally, we found that LS cells with reward proximate firing are more highly cross correlated with hippocampal firing than cells without reward proximate place fields (Figure 7). This suggests that a predominant input to these LS cells that are biased to reward locations is from the hippocampus.

In the present task, the probability of finding a HPC place field increases as the animal approaches final goal location (Figure 6H). However, the representation of the goal is decreased relative to locations proximate to the forced and choice points of the maze. Past work has found that the HPC uniquely over-represents salient or goal locations on tasks demanding increased spatial memory (Dupret et al., 2010). It is possible, therefore, since the important spatial memory components of this task occurred at the forced and choice points, that these locations came to be even more over-represented in the HPC that the goal locations.

Reward-related spatial information may be transmitted from the HPC to the LS in multiple ways. First, there may be a fixed population of HPC cells that consistently represent reward, which selectively innervate the LS. There is direct evidence for such a population (Gauthier and Tank, 2018), but it is unknown if it comprises a large portion of LS afferents. Alternatively, different HPC place cells may represent reward in different spatial environments, but the LS cells receiving place cell input are differentially modulated by external inputs depending on the specific task.

So how are reward proximate place cells more common in the LS than in the hippocampus? First, it is possible that hippocampal place cells that have fields proximate to reward locations selectively, either anatomically or functionally, innervate LS cells (Figure 7D, Model 1). Additionally, it is possible that non-reward related place cell inputs from the hippocampus are diluted in the LS if they converge with non-reward related input. This input may be from other place cells without reward proximate fields (Figure 7D, Model 2), or from other areas, such as speed or acceleration related information from the brainstem (Figure 7D, model 3; Wirtshafter and Wilson, 2019). In this model, reward proximate place cells would innervate a separate group of LS cells which would account for the over-representation of reward proximate place fields in the LS (Figure 7D, Model 2). Our finding that LS and HPC place cells with fields proximate to reward have increased cross correlations compared to cells with more distal place fields is consistent with any of these models, as the higher cross correlation may be due to a selective innervation of the LS by reward-proximate HPC cells, or the ‘dilution’ of non-reward related place cells by competing inputs (Figure 7D, Model 3). However, our previous work (Wirtshafter and Wilson, 2019) has shown that LS cells containing place information are less likely to contain speed or acceleration information than cells without place fields, suggesting that speed or acceleration information may be competing with spatial information sent from the HPC, a result most easily accommodated by model 3.

We also found differences in place cell activity based on whether the animal was moving toward or away from the reward site. We first found that the reward site itself was overrepresented in LS place fields right after the reward site had been visited (Figure 6G). We also found that during reward approach, within a place field, LS place cells, on average ramped up firing (Figure 5E). It is possible that both this overrepresentation and difference in skew may serve memory or planning functions during navigation. We have previously found that the HPC-LS coherence increased during working memory tasks (Wirtshafter and Wilson, 2019), and increased place field activity at a site of significance could be evidence of hippocampal and LS coordination during encoding of significant task locations. There is increased VTA reactivation of reward related cells during ripples and VTA cells also ramp up firing during reward approach (Gomperts et al., 2015), showing that the significance of reward related locations has been established downstream of the LS. That information may be communicated to the VTA from the LS by the over-representation of just-visited reward related locations in place cell firing, as well as by increased within-field firing upon reward approach.

The distribution of field numbers was not significantly different between HPC and CA1 cells (two sample Kolmogorov-Smirnov (KS) test, p>0.5).

Top panel is traveling to reward, bottom panel is traveling away. Because of the non-linearity of the track, different place fields were visited different numbers of times, and therefore each field contributes a different number of bins to these histograms.

Data has been deposited to Collaborative Research in Computational Neuroscience (CRNRS) under the accession code hc-29 (https://doi.org/10.6080/K0NG4NV8). Users must first create a free account (https://crcns.org/register) before they can download the datasets from the site. All analysis code is available at https://github.com/hsw28/data_analysis (copy archived at https://github.com/elifesciences-publications/data_analysis).

1. 1. Wirtshafter HS
   2. Wilson MA

   (2020) Collaborative Research in Computational Neuroscience

   Tetrode recordings of hippocampus CA1 and dorsal lateral septum in rat.

   https://doi.org/10.6080/K0NG4NV8

1. 1. Bezzi M
   2. Samengo I
   3. Leutgeb S
   4. Mizumori SJ

   (2002) Measuring information spatial densities

   Neural Computation 14:405–420.

   https://doi.org/10.1162/08997660252741176

   • PubMed
   • Google Scholar
2. 1. Bezzi M

   (2005) Information maximization for exploring neural coding in Hippocampus and lateral septum

   Biosystems 79:183–189.

   https://doi.org/10.1016/j.biosystems.2004.09.019

   • PubMed
   • Google Scholar
3. 1. Dupret D
   2. O'Neill J
   3. Pleydell-Bouverie B
   4. Csicsvari J

   (2010) The reorganization and reactivation of hippocampal maps predict spatial memory performance

   Nature Neuroscience 13:995–1002.

   https://doi.org/10.1038/nn.2599

   • PubMed
   • Google Scholar
4. 1. Ego-Stengel V
   2. Wilson MA

   (2007) Spatial selectivity and theta phase precession in CA1 interneurons

   Hippocampus 17:161–174.

   https://doi.org/10.1002/hipo.20253

   • Google Scholar
5. 1. Fenton AA
   2. Kao HY
   3. Neymotin SA
   4. Olypher A
   5. Vayntrub Y
   6. Lytton WW
   7. Ludvig N

   (2008) Unmasking the CA1 ensemble place code by exposures to small and large environments: more place cells and multiple, irregularly arranged, and expanded place fields in the larger space

   Journal of Neuroscience 28:11250–11262.

   https://doi.org/10.1523/JNEUROSCI.2862-08.2008

   • PubMed
   • Google Scholar
6. 1. Frank LM
   2. Stanley GB
   3. Brown EN

   (2004) Hippocampal plasticity across multiple days of exposure to novel environments

   Journal of Neuroscience 24:7681–7689.

   https://doi.org/10.1523/JNEUROSCI.1958-04.2004

   • PubMed
   • Google Scholar
7. 1. Fyhn M
   2. Molden S
   3. Witter MP
   4. Moser EI
   5. Moser MB

   (2004) Spatial representation in the entorhinal cortex

   Science 305:1258–1264.

   https://doi.org/10.1126/science.1099901

   • PubMed
   • Google Scholar
8. 1. Fyhn M
   2. Hafting T
   3. Treves A
   4. Moser MB
   5. Moser EI

   (2007) Hippocampal remapping and grid realignment in entorhinal cortex

   Nature 446:190–194.

   https://doi.org/10.1038/nature05601

   • PubMed
   • Google Scholar
9. 1. Gauthier JL
   2. Tank DW

   (2018) A dedicated population for reward coding in the Hippocampus

   Neuron 99:179–193.

   https://doi.org/10.1016/j.neuron.2018.06.008

   • PubMed
   • Google Scholar
10. 1. Gomperts SN
    2. Kloosterman F
    3. Wilson MA

    (2015) VTA neurons coordinate with the hippocampal reactivation of spatial experience

    eLife 4:e05360.

    https://doi.org/10.7554/eLife.05360

    • PubMed
    • Google Scholar
11. 1. Hafting T
    2. Fyhn M
    3. Molden S
    4. Moser MB
    5. Moser EI

    (2005) Microstructure of a spatial map in the entorhinal cortex

    Nature 436:801–806.

    https://doi.org/10.1038/nature03721

    • PubMed
    • Google Scholar
12. Software

    1. Hale G
    2. Wirtshaftee HS

    (2019) arte-backend, version v1.0

    Github.

    https://github.com/hsw28/arte-hardware
13. 1. Hasselmo ME
    2. Eichenbaum H

    (2005) Hippocampal mechanisms for the context-dependent retrieval of episodes

    Neural Networks 18:1172–1190.

    https://doi.org/10.1016/j.neunet.2005.08.007

    • Google Scholar
14. 1. Hollup SA
    2. Molden S
    3. Donnett JG
    4. Moser MB
    5. Moser EI

    (2001) Accumulation of hippocampal place fields at the goal location in an annular watermaze task

    The Journal of Neuroscience 21:1635–1644.

    https://doi.org/10.1523/JNEUROSCI.21-05-01635.2001

    • PubMed
    • Google Scholar
15. 1. Ji D
    2. Wilson MA

    (2007) Coordinated memory replay in the visual cortex and Hippocampus during sleep

    Nature Neuroscience 10:100–107.

    https://doi.org/10.1038/nn1825

    • PubMed
    • Google Scholar
16. 1. Jiang JX
    2. Liu H
    3. Huang ZZ
    4. Cui Y
    5. Zhang XQ
    6. Zhang XL
    7. Cui Y
    8. Xin WJ

    (2018) The role of CA3-LS-VTA loop in the formation of conditioned place preference induced by context-associated reward memory for morphine

    Addiction Biology 23:41–54.

    https://doi.org/10.1111/adb.12468

    • PubMed
    • Google Scholar
17. 1. Jones MW
    2. Wilson MA

    (2005a) Phase precession of medial prefrontal cortical activity relative to the hippocampal theta rhythm

    Hippocampus 15:867–873.

    https://doi.org/10.1002/hipo.20119

    • PubMed
    • Google Scholar
18. 1. Jones MW
    2. Wilson MA

    (2005b) Theta rhythms coordinate hippocampal-prefrontal interactions in a spatial memory task

    PLOS Biology 3:e402.

    https://doi.org/10.1371/journal.pbio.0030402

    • PubMed
    • Google Scholar
19. 1. Kita T
    2. Nishijo H
    3. Eifuku S
    4. Terasawa K
    5. Ono T

    (1995) Place and contingency differential responses of monkey septal neurons during conditional place-object discrimination

    The Journal of Neuroscience 15:1683–1703.

    https://doi.org/10.1523/JNEUROSCI.15-03-01683.1995

    • PubMed
    • Google Scholar
20. 1. Knierim JJ

    (2002) Dynamic interactions between local surface cues, distal landmarks, and intrinsic circuitry in hippocampal place cells

    The Journal of Neuroscience 22:6254–6264.

    https://doi.org/10.1523/JNEUROSCI.22-14-06254.2002

    • PubMed
    • Google Scholar
21. 1. Knierim JJ
    2. Neunuebel JP
    3. Deshmukh SS

    (2014) Functional correlates of the lateral and medial entorhinal cortex: objects, path integration and local–global reference frames

    Philosophical Transactions of the Royal Society B: Biological Sciences 369:20130369.

    https://doi.org/10.1098/rstb.2013.0369

    • Google Scholar
22. 1. Knierim JJ
    2. Hamilton DA

    (2011) Framing spatial cognition: neural representations of proximal and distal frames of reference and their roles in navigation

    Physiological Reviews 91:1245–1279.

    https://doi.org/10.1152/physrev.00021.2010

    • PubMed
    • Google Scholar
23. 1. Kobayashi T
    2. Nishijo H
    3. Fukuda M
    4. Bures J
    5. Ono T

    (1997) Task-dependent representations in rat hippocampal place neurons

    Journal of Neurophysiology 78:597–613.

    https://doi.org/10.1152/jn.1997.78.2.597

    • PubMed
    • Google Scholar
24. 1. Kobayashi T
    2. Tran AH
    3. Nishijo H
    4. Ono T
    5. Matsumoto G

    (2003) Contribution of hippocampal place cell activity to learning and formation of goal-directed navigation in rats

    Neuroscience 117:1025–1035.

    https://doi.org/10.1016/S0306-4522(02)00700-5

    • PubMed
    • Google Scholar
25. 1. Krettek JE
    2. Price JL

    (1978) Amygdaloid projections to subcortical structures within the basal forebrain and brainstem in the rat and cat

    The Journal of Comparative Neurology 178:225–253.

    https://doi.org/10.1002/cne.901780204

    • PubMed
    • Google Scholar
26. 1. Lee I
    2. Rao G
    3. Knierim JJ

    (2004) A double dissociation between hippocampal subfields: differential time course of CA3 and CA1 place cells for processing changed environments

    Neuron 42:803–815.

    https://doi.org/10.1016/j.neuron.2004.05.010

    • PubMed
    • Google Scholar
27. 1. Lee I
    2. Griffin AL
    3. Zilli EA
    4. Eichenbaum H
    5. Hasselmo ME

    (2006) Gradual translocation of spatial correlates of neuronal firing in the Hippocampus toward prospective reward locations

    Neuron 51:639–650.

    https://doi.org/10.1016/j.neuron.2006.06.033

    • PubMed
    • Google Scholar
28. 1. Leutgeb S
    2. Mizumori SJY

    (2002) Context-specific spatial representations by lateral septal cells

    Neuroscience 112:655–663.

    https://doi.org/10.1016/S0306-4522(02)00101-X

    • Google Scholar
29. 1. Luo AH
    2. Tahsili-Fahadan P
    3. Wise RA
    4. Lupica CR
    5. Aston-Jones G

    (2011) Linking context with reward: a functional circuit from hippocampal CA3 to ventral tegmental area

    Science 333:353–357.

    https://doi.org/10.1126/science.1204622

    • PubMed
    • Google Scholar
30. 1. Lyttle D
    2. Gereke B
    3. Lin KK
    4. Fellous J-M

    (2013) Spatial scale and place field stability in a grid-to-place cell model of the dorsoventral axis of the hippocampus

    Hippocampus 23:729–744.

    https://doi.org/10.1002/hipo.22132

    • Google Scholar
31. 1. Markus EJ
    2. Barnes CA
    3. McNaughton BL
    4. Gladden VL
    5. Skaggs WE

    (1994) Spatial information content and reliability of hippocampal CA1 neurons: effects of visual input

    Hippocampus 4:410–421.

    https://doi.org/10.1002/hipo.450040404

    • PubMed
    • Google Scholar
32. 1. Mathieu-Kia AM
    2. Pages C
    3. Besson MJ

    (1998) Inducibility of c-Fos protein in visuo-motor system and limbic structures after acute and repeated administration of nicotine in the rat

    Synapse 29:343–354.

    https://doi.org/10.1002/(SICI)1098-2396(199808)29:4<343::AID-SYN6>3.0.CO;2-5

    • PubMed
    • Google Scholar
33. 1. McGlinchey EM
    2. Aston-Jones G

    (2018) Dorsal Hippocampus drives Context-Induced cocaine seeking via inputs to lateral septum

    Neuropsychopharmacology 43:987–1000.

    https://doi.org/10.1038/npp.2017.144

    • PubMed
    • Google Scholar
34. 1. Mehta MR
    2. Barnes CA
    3. McNaughton BL

    (1997) Experience-dependent, asymmetric expansion of hippocampal place fields

    PNAS 94:8918–8921.

    https://doi.org/10.1073/pnas.94.16.8918

    • PubMed
    • Google Scholar
35. 1. Mehta MR
    2. Quirk MC
    3. Wilson MA

    (2000) Experience-dependent asymmetric shape of hippocampal receptive fields

    Neuron 25:707–715.

    https://doi.org/10.1016/S0896-6273(00)81072-7

    • PubMed
    • Google Scholar
36. 1. Monaco JD
    2. Knierim JJ
    3. Zhang K

    (2011) Sensory feedback, error correction, and remapping in a multiple oscillator model of place-cell activity

    Frontiers in Computational Neuroscience 5:39.

    https://doi.org/10.3389/fncom.2011.00039

    • PubMed
    • Google Scholar
37. 1. Monaco JD
    2. De Guzman RM
    3. Blair HT
    4. Zhang K

    (2019) Spatial synchronization codes from coupled rate-phase neurons

    PLOS Computational Biology 15:e1006741.

    https://doi.org/10.1371/journal.pcbi.1006741

    • PubMed
    • Google Scholar
38. 1. Moser EI
    2. Kropff E
    3. Moser MB

    (2008) Place cells, grid cells, and the brain's spatial representation system

    Annual Review of Neuroscience 31:69–89.

    https://doi.org/10.1146/annurev.neuro.31.061307.090723

    • PubMed
    • Google Scholar
39. 1. Nerad L
    2. McNaughton N

    (2006) The septal EEG suggests a distributed organization of the pacemaker of hippocampal theta in the rat

    European Journal of Neuroscience 24:155–166.

    https://doi.org/10.1111/j.1460-9568.2006.04902.x

    • PubMed
    • Google Scholar
40. Software

    1. Newman J
    2. Hale G
    3. Myroshnychenko M
    4. Voigts J
    5. Flores FJ
    6. DonatoRidgley I

    (2017)  jonnew/Oat: Oat Version 1.0

    Zenodo.

    http://doi.org/10.5281/zenodo.1098579
41. 1. O'Keefe J
    2. Dostrovsky J

    (1971) The Hippocampus as a spatial map preliminary evidence from unit activity in the freely-moving rat

    Brain Research 34:171–175.

    https://doi.org/10.1016/0006-8993(71)90358-1

    • PubMed
    • Google Scholar
42. Book

    1. O'Keefe J
    2. Nadel L

    (1978) The Hippocampus as a Cognitive Map

    Oxford University Press.

    https://doi.org/10.1017/S0140525X00063949

    • Google Scholar
43. 1. Olds J
    2. Milner P

    (1954) Positive reinforcement produced by electrical stimulation of septal area and other regions of rat brain

    Journal of Comparative and Physiological Psychology 47:419–427.

    https://doi.org/10.1037/h0058775

    • PubMed
    • Google Scholar
44. 1. Olton DS
    2. Walker JA
    3. Gage FH

    (1978) Hippocampal connections and spatial discrimination

    Brain Research 139:295–308.

    https://doi.org/10.1016/0006-8993(78)90930-7

    • PubMed
    • Google Scholar
45. 1. Oshima H
    2. Katayama Y

    (2010) Neuroethics of deep brain stimulation for mental disorders: brain stimulation reward in humans

    Neurologia Medico-Chirurgica 50:845–852.

    https://doi.org/10.2176/nmc.50.845

    • PubMed
    • Google Scholar
46. 1. Quirk GJ
    2. Muller RU
    3. Kubie JL
    4. Ranck JB

    (1992) The positional firing properties of medial entorhinal neurons: description and comparison with hippocampal place cells

    The Journal of Neuroscience 12:1945–1963.

    https://doi.org/10.1523/JNEUROSCI.12-05-01945.1992

    • PubMed
    • Google Scholar
47. 1. Rich PD
    2. Liaw HP
    3. Lee AK

    (2014) Place cells large environments reveal the statistical structure governing hippocampal representations

    Science 345:814–817.

    https://doi.org/10.1126/science.1255635

    • PubMed
    • Google Scholar
48. 1. Risold PY
    2. Swanson LW

    (1997) Connections of the rat lateral septal complex

    Brain Research. Brain Research Reviews 24:115–195.

    https://doi.org/10.1016/s0165-0173(97)00009-x

    • PubMed
    • Google Scholar
49. 1. Roesch MR
    2. Calu DJ
    3. Schoenbaum G

    (2007) Dopamine neurons encode the better option in rats deciding between differently delayed or sized rewards

    Nature Neuroscience 10:1615–1624.

    https://doi.org/10.1038/nn2013

    • PubMed
    • Google Scholar
50. 1. Sheehan TP
    2. Chambers RA
    3. Russell DS

    (2004) Regulation of affect by the lateral septum: implications for neuropsychiatry

    Brain Research Reviews 46:71–117.

    https://doi.org/10.1016/j.brainresrev.2004.04.009

    • PubMed
    • Google Scholar
51. 1. Siegle JH
    2. Wilson MA

    (2014) Enhancement of encoding and retrieval functions through theta phase-specific manipulation of Hippocampus

    eLife 3:e03061.

    https://doi.org/10.7554/eLife.03061

    • PubMed
    • Google Scholar
52. 1. Skaggs WE
    2. McNaughton BL

    (1998) Spatial firing properties of hippocampal CA1 populations in an environment containing two visually identical regions

    The Journal of Neuroscience 18:8455–8466.

    https://doi.org/10.1523/JNEUROSCI.18-20-08455.1998

    • PubMed
    • Google Scholar
53. 1. Sotomayor R
    2. Forray MI
    3. Gysling K

    (2005) Acute morphine administration increases extracellular DA levels in the rat lateral septum by decreasing the GABAergic inhibitory tone in the ventral tegmental area

    Journal of Neuroscience Research 81:132–139.

    https://doi.org/10.1002/jnr.20537

    • PubMed
    • Google Scholar
54. Book

    1. Spiegel MR

    (1961)

    Schaum's Outline of Theory and Problems of Statistics

    New York: Schaum Pub Co.

    • Google Scholar
55. Conference

    1. Swanson LW

    (1977)

    The anatomical organization of septo-hippocampal projections

    Ciba Foundation Symposium. pp. 25–48.

    • Google Scholar
56. 1. Swanson LW
    2. Sawchenko PE
    3. Cowan WM

    (1981) Evidence for collateral projections by neurons in Ammon's horn, the dentate gyrus, and the subiculum: a multiple retrograde labeling study in the rat

    The Journal of Neuroscience 1:548–559.

    https://doi.org/10.1523/JNEUROSCI.01-05-00548.1981

    • PubMed
    • Google Scholar
57. 1. Sweeney P
    2. Yang Y

    (2015) An excitatory ventral Hippocampus to lateral septum circuit that suppresses feeding

    Nature Communications 6:10188.

    https://doi.org/10.1038/ncomms10188

    • PubMed
    • Google Scholar
58. 1. Sweeney P
    2. Yang Y

    (2016) An inhibitory septum to lateral hypothalamus circuit that suppresses feeding

    The Journal of Neuroscience 36:11185–11195.

    https://doi.org/10.1523/JNEUROSCI.2042-16.2016

    • PubMed
    • Google Scholar
59. 1. Takamura Y
    2. Tamura R
    3. Zhou TL
    4. Kobayashi T
    5. Tran AH
    6. Eifuku S
    7. Ono T

    (2006) Spatial firing properties of lateral septal neurons

    Hippocampus 16:635–644.

    https://doi.org/10.1002/hipo.20196

    • PubMed
    • Google Scholar
60. 1. Tingley D
    2. Buzsáki G

    (2018) Transformation of a spatial map across the Hippocampal-Lateral septal circuit

    Neuron 98:1229–1242.

    https://doi.org/10.1016/j.neuron.2018.04.028

    • PubMed
    • Google Scholar
61. 1. Wilson MA
    2. McNaughton BL

    (1994) Reactivation of hippocampal ensemble memories during sleep

    Science 265:676–679.

    https://doi.org/10.1126/science.8036517

    • PubMed
    • Google Scholar
62. Software

    1. Wirtshafter HS

    (2020) Data analysis for 'Differences in reward biased spatialrepresentations in the lateral septum andhippocampus', version 5315fb2

    GitHub.

    https://github.com/hsw28/data_analysis
63. 1. Wirtshafter HS
    2. Wilson MA

    (2019) Locomotor and hippocampal processing converge in the lateral septum

    Current Biology 29:3177–3192.

    https://doi.org/10.1016/j.cub.2019.07.089

    • PubMed
    • Google Scholar
64. 1. Zhou TL
    2. Tamura R
    3. Kuriwaki J
    4. Ono T

    (1999) Comparison of medial and lateral septal neuron activity during performance of spatial tasks in rats

    Hippocampus 9:220–234.

    https://doi.org/10.1002/(SICI)1098-1063(1999)9:3<220::AID-HIPO3>3.0.CO;2-E

    • PubMed
    • Google Scholar
65. 1. Zhou Y
    2. Zhu H
    3. Liu Z
    4. Chen X
    5. Su X
    6. Ma C
    7. Tian Z
    8. Huang B
    9. Yan E
    10. Liu X
    11. Ma L

    (2019) A ventral CA1 to nucleus accumbens core engram circuit mediates conditioned place preference for cocaine

    Nature Neuroscience 22:1986–1999.

    https://doi.org/10.1038/s41593-019-0524-y

    • PubMed
    • Google Scholar

Author details

1. Hannah S Wirtshafter

   1. Department of Biology, MassachusettsInstitute of Technology, Cambridge, United States
   2. Picower Institute for Learning andMemory, Massachusetts Institute of Technology, Cambridge, United States

   Contribution

   Conceptualization, Data curation, Formal analysis, Funding acquisition, Validation, Investigation, Methodology, Writing - original draft, Writing - review and editing

   For correspondence

   hsw@mit.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-4684-7074

2. Matthew A Wilson

   1. Department of Biology, MassachusettsInstitute of Technology, Cambridge, United States
   2. Picower Institute for Learning andMemory, Massachusetts Institute of Technology, Cambridge, United States
   3. Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, Cambridge, United States

   Contribution

   Supervision, Funding acquisition, Project administration, Writing - review and editing

   Competing interests

   No competing interests declared

• 3,310

  views

• 472

  downloads

• 37

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.