Learning is a process of acquiring new knowledge required for appropriate behavior and is highly dependent on our previous experience. Our brain integrates incoming sensory information with history information of previous stimuli to form a knowledgeable association of the current stimulus. Although the strong link between history (i.e., past experience) and learning, the underlying cortex-wide dynamics are relatively unknown, partially because most previous studies separately focus either on learning or history (Hattori et al., 2019). Learning-related neuronal dynamics are broadly observed across the whole cortex, including primary sensory and motor areas (Blake et al., 2002; Chen et al., 2015; Gilad and Helmchen, 2020; Jurjut et al., 2017; Komiyama et al., 2010; Li et al., 2008; Poort et al., 2015; Makino and Komiyama, 2015; Yan et al., 2014), higher-order association areas (Driscoll et al., 2017; Gilad and Helmchen, 2020), and prefrontal cortex (Le Merre et al., 2018; Pasupathy and Miller, 2005). But do these areas that participate in the learning process also carry trial history information?

Encoding of information carried on from the previous trial (i.e., trial history) has been reported mainly in higher-order cortical areas such as the posterior parietal cortex (PPC) (Akrami et al., 2018; Harvey et al., 2012; Hwang et al., 2017; Morcos and Harvey, 2016; Scott et al., 2017; Suzuki et al., 2022), retrosplenial cortex (Hattori et al., 2019; Vann et al., 2009), and prefrontal cortex (Banerjee et al., 2020; Johnson et al., 2016; Kawai et al., 2015; Scott et al., 2017; Sul et al., 2010; Tsutsui et al., 2016), and to a smaller extent in lower-order primary sensory areas such as BC (Banerjee et al., 2020; Chéreau et al., 2020; Rodgers et al., 2021). It is still unknown on how different cortical areas encode trial history with regard to learning. In other words, does trial history encoding in the cortex change as a function of learning? Another important aspect of the history learning is the temporal relationship between trial history encoding and the current stimulus. For example, does trial history emerge in cortex before the current incoming stimulus, or maybe both past and present information emerge simultaneously in a certain cortical area? From the temporal aspect, optogenetic silencing of PPC area during the inter-trial interval affected performance, highlighting that higher-order cortical areas may maintain history information before the incoming current stimulus (Akrami et al., 2018; Hwang et al., 2017).

To study the history-learning relationship, we use wide-field cortical imaging of mice learning to discriminate between two textures and focus on the cortex-wide dynamics of trial history. In a previous study using the same dataset, we showed that mice learning a whisker-based texture discrimination task, increase activity in task-related areas (e.g., BC and rostrolateral association cortex [RL]) as they become experts (Gilad and Helmchen, 2020). RL is part of the PPC and is located within the cluster of higher-order association areas surrounding V1. RL plays pivotal roles in cross-modal sensory integration, learning, and history, but the relationship between history and learning in RL is relatively unknown (Akrami et al., 2018; Driscoll et al., 2017; Hattori et al., 2019; Hwang et al., 2017; Khodagholy et al., 2017; Morcos and Harvey, 2016; Save and Poucet, 2009). By classifying trials according to the preceding trial, we were able to detect trial history information that emerges only as the mouse gains expertise. Specifically, trial history emerges in RL, before stimulus presentation and then is transferred to BC during stimulus presentation, which may aid in learning the rewarded stimulus.

In this study, we investigate trial history dynamics across the whole dorsal cortex and its emergence during learning in transgenic mice expressing a calcium indicator (GCaMP6f) in L2/3 excitatory neurons (n = 7 mice). This dataset is identical to the one published in Gilad and Helmchen, 2020 where we focused only on learning dynamics. Using wide-field calcium imaging through the intact skull (Gallero-Salas et al., 2021; Gilad et al., 2018; Gilad and Helmchen, 2020; Vanni and Murphy, 2014), we chronically measured large-scale neocortical L2/3 activity in the contralateral hemisphere as mice learned a go/no-go whisker-dependent texture discrimination task (Gilad and Helmchen, 2020). Whisker movements and body movements were video monitored and synchronized to the calcium imaging data (Materials and methods). To map the dorsal cortex, we functionally mapped sensory areas for each mouse during anesthesia (see Materials and methods). Based on these maps (and skull coordinates) we registered all images to the 2D topview Allen reference atlas (Oh et al., 2014) and defined 25 areas of interest, further divided into four groups (Figure 1c; Gilad and Helmchen, 2020).

Figure 1 with 3 supplements see all

Download asset Open asset

Mice were trained on a head-fixed, whisker-based go/no-go texture discrimination task (Chen et al., 2013; Gilad and Helmchen, 2020; Figure 1a; Materials and methods). Each trial started with an auditory cue (stimulus cue), signaling the approach of either two types of sandpapers (grit size P100: rough texture; P1200: smooth texture; 3M) to the mouse’s whiskers as ‘go’ or ‘no-go’ textures. The texture stayed in touch with the whiskers for 2 s, and then it was moved out after which an additional auditory cue (response cue) signaled the start of a 2-s response period (Figure 1b) followed by a 6-s break until the next trial auditory cue. Five mice were trained to lick for the P100 and two mice were trained to lick for the P1200 texture. Mice were rewarded in ‘Hit’ trials for correctly licking after the go texture and punished with white noise for incorrectly licking for the no-go texture (‘false alarm’ trials, FA). Mice were neither rewarded nor punished when they withheld licking for the go and no-go textures (‘Miss’ and ‘correct-rejection’, CR, trials, respectively). We defined two time windows within the trial structure: the ‘pre period’ when the texture approaches the whiskers (−1 to −0.6 s relative to the texture stop; mainly before the first whisker-texture touch); and the ‘stim period’ during texture touch (−0.2 to 0.2 s relative to texture stop; Figure 1b).

The performance of all mice increased with training (5–11 days; ~500 trials/day) and eventually reached high discrimination levels (quantified by d′; Figure 1—figure supplement 1; Gilad et al., 2018; Materials and methods). We defined the ‘learning threshold’ of reaching expert level for each mouse by crossing the inflection point of the sigmoid fit for the learning curve (in units of ‘trial number’; Figure 1e, Figure 1—figure supplement 1). The fastest learning mouse reached threshold in slightly less than thousand trials whereas mouse #4 took substantially longer (Figure 1—figure supplement 1). In addition, we defined a naive (first day of recording), learning (day of crossing the learning threshold; second or third day), and expert (last recording day) phases for each mouse (Figure 1e). All mice, after gaining expertise, showed strong activation in the BC (Figure 1d, upper panel). This activation was during stimulus representation, stronger in Hit trials compared to CR trials (Figure 1d, lower panel), not dependent on the texture type (i.e., if the hit was P100 or P1200).

Here, we focus on the trial history content for each trial type. We subgrouped all the Hit trials (i.e., the current trial type) based on the previous trial type (i.e., trial history): CR (‘CR-Hit’; n = 423 ± 74, mean ± standard error of the mean [SEM]), Hit (‘Hit-Hit’; n = 585 ± 42), FA (‘FA-Hit’; n = 217 ± 24), or Miss (‘Miss-Hit’; n = 55 ± 24; Figure 1f, g). ‘Miss-Hit’ were not analyzed due to a small number of trials. Our main analysis will compare ‘CR-Hit’ (orange) and ‘Hit-Hit’ (blue) trial pairs, since they are present in large numbers during all phases in each mouse separately (Figure 1g; but see Figure 2—figure supplement 7 for a comparison of other trial pairs). We note that learning curves that are calculated separately for each pair (i.e., either a preceding Hit or CR trial) were not significantly different (Figure 1—figure supplement 2). We further note FA probabilities did not significantly differ based on the preceding trial type (Figure 1—figure supplement 3). In addition, the lick reaction time (but not the lick rate) between Hit-Hit and CR-Hit were significantly different (p < 0.05; Wilcoxon signed-rank test). We emphasize that in this comparison, the current trial type is identical (i.e., Hit) whereas only the pervious trial (i.e., the history, CR, or Hit) differed, therefore eliminating activity differences due to the current stimulus.

Trial history in BC emerges during learning

First, we focused on trial history encoding in BC, specifically during the stim period. BC displayed higher activity during CR-Hit compared to Hit-Hit only during learning and expert phases, but not during the naive period (Figure 2a, Figure 2—figure supplement 1). This difference was significant during the stim period in learning and expert phases across mice (Figure 2b; two-way analysis of variance [ANOVA] with repeated measures; DF(1-6) F = 51 p < 0.001, DF(2-12) F = 18 p < 0.001, DF(2-12) F = 5 p < 0.05 for trial history, learning, and the interaction between trial history and learning; post hoc Tukey analysis p < 0.05 for trial history in learning and expert phases; p > 0.05 in the naive phase). In addition, a statistical comparison between CR-Hit and Hit-Hit responses within each mouse separately maintained significance for expert (7/7 mice; Mann–Whitney U-test p < 0.05) and learning (6/7 mice) but not for naive (0/7 mice) (Figure 2—figure supplement 1). We further report that responses during the reward period in cortex and specifically in BC went back to baseline 4–5 s after the start of the reward period and 6–8 s before the presentation of the next stimulus (total inter-trial interval ranged between 10 and 12 s). In addition, responses in BC during the reward period were not consistently modulated as a function of learning (p > 0.05; Wilcoxon signed-rank test between naive and expert, BC response averaged during the reward period, 2–4 s after stimulus onset; n = 7 mice). Taken together, we find that direct responses from the reward period do not affect history-related responses during the next trial.

Figure 2 with 7 supplements see all

Download asset Open asset

To control for the possible contamination of non-calcium-related signals such as hemodynamics, we performed a battery of additional experiments (see Materials and methods). First, two mice performed the task (expert) while we excited the cortex with a control light (510 nm, isosbestic wavelength). Correcting the original signal with the control light maintained significant difference in trial history (Figure 2—figure supplement 2). Second, we trained an additional three mice on the same task and imaged their cortex using an interleaved protocol of 473 (calcium signal) and 405 (control signal, isosbestic wavelength) nm lights. The corrected signal (473 signal minus the 405 signal; see Materials and methods) maintained a significant trial history difference between CR-Hit and Hit-Hit conditions during learning and expert phases, but not during the naive phase (Figure 2—figure supplement 3). Finally, we further performed two photon imaging of single cells in BC and RL during the expert phase and found significant trial history differences, that is, higher response in CR-Hit compared to Hit-Hit in BC (during the stim period) and RL (Figure 2—figure supplement 4; during the pre period). Taken together, non-calcium dynamics such as hemodynamics have a minor effect the results, specifically regarding trial history differences.

To check whether this effect is not due to difference in body or whisker movements between the two pair types, we calculated the body movements (1 − frame-to-frame correlation in mouth, forelimb and hindlimb areas) and whisker envelope as a function of time (see Materials and methods). Both body movements and whisker envelope were similar between CR-Hit and Hit-Hit pairs (Figure 2c) and there was no significant difference across mice during the stim period for neither naive, learning, or expert phases (Figure 2d, p > 0.05; signed-rank test) nor during the pre period (p > 0.05, signed-rank test, data not shown). A within mouse statistical comparison between body or whisker parameters in CR-Hit and Hit-Hit maintained a non-significant difference in expert (1/7 mice were significantly different; Mann–Whitney U-test p > 0.05), learning (2/7 mice) and naive (0/7 mice). In addition, we performed a more detailed body and whisker analysis, for example, decomposing the movement to different body parts and obtaining single whisker dynamics. These analyses did not find significant differences in movement parameters between CR-Hit and Hit-Hit conditions (Figure 2—figure supplement 5 and Figure 2—figure supplement 6). These results, along with the fact that the current trial type in both conditions is identical, strongly indicate the presence of trial history information in BC.

We next quantified the emergence of trial history with regard to the different time scales, the trial structure (within seconds) or the learning profile (across days). We first show 2D activity plots in BC for each trial pair (i.e., CR-Hit and Hit-Hit; showing activity of only the Hit trial), where trial time is plotted on the x-axis and trial number across learning time on the y-axis (Figure 3a; 100-trial bins regardless of trial pair). Both trial pairs display an increase in activity during the stim period slightly after passing the learning threshold. We defined a history modulation index as the difference in activity for BC between the two pair types (Hit-CR minus Hit-Hit). History modulation increased around the stim period only in learning and expert phases but not in the naive case (Figure 3b, c). A significant history modulation was defined as values exceeding mean ± 2 standard deviation (SD) of a trial-shuffled sample distribution (n = 1000 iterations) and was performed for each mouse separately (Figure 3b). The onset of the history modulation was defined as the first-time frame reaching significant values (red arrows in Figure 3b) and was found in BC within the stim period (Figure 3d; 0.08 ± 0.28 s, −0.32 ± 0.28 s, median ± SEM relative to texture stop in learning and expert phases, respectively). We note that in the expert phase there is also a small peak exceeding the significance around the cue, indicating history information in BC may be present to some extent before stimulus presentation. Next, we quantified the history modulation in BC during the stim period as a function of the learning profile. History modulation in BC had the steepest increase after mice crossed their learning threshold (Figure 3e, f). The onset of the history modulation was defined as the first trial bin exceeding mean ± 2 standard deviation of trial-shuffled sample distribution and was found to occur shortly after the learning threshold, highly correlated with the learning threshold indicating strong relationship between history emergence and learning of each individual mouse (Figure 3g, h; 500 ± 221 trials, median ± SEM, r = 0.97, p < 0.001, Spearman correlation). Note that our definition of significance is relatively strict and an increase in history information can be observed shortly (i.e., tens of trials) after crossing the learning threshold (Figure 3e).

Figure 3

Download asset Open asset

We expanded our trial history analysis also for the other pair types other than CR-Hit and Hit-Hit. For sufficient trial numbers, we focused on the learning phase. First, we compare FA-Hit to Hit-Hit and CR-Hit, that is, the same current trial type but preceded by an error trial (FA). Response in BC for FA-Hit was similar to Hit-Hit and significantly lower compared to CR-Hit (Figure 2—figure supplement 7; p < 0.05 signed-rank test). This result highlights that specifically a correct rejection (CR), rather than the stimulus (i.e., texture) type, has a strong history effect. Next, we compared FA-CR, Hit-CR, and CR-CR, that is, similar to the previous comparison differing only in the current trial type (CR instead of Hit). There was no significant difference between the different pairs, indicating that the current trial type, that is, Hit in this case, has a strong effect along with the history of the CR (Figure 2—figure supplement 7; p > 0.05, signed-rank test). A comparison of FA-FA, Hit-FA, and CR-FA did not show a significant difference (Figure 2—figure supplement 7; p > 0.05, signed-rank test). In general, a preceding CR trial resulted in higher activation independent of the current trial type (i.e., Hit, CR, or FA; not significant for CR and FA), indicating that history information is present at the current time independently of incoming sensory information (Figure 2—figure supplement 7; compare orange bars to the blue bars). In conclusion, we found that the CR-Hit pair displayed a specific enhancement in BC which is related both to the preceding and current trial type (see Discussion).

Next, we expanded our analysis to the whole dorsal cortex during the stim period. Mean activation maps for both CR-Hit and Hit-Hit pairs (i.e., activity for the current Hit trial whereas only the preceding trial was different) during the stim period displayed a pronounced activation patch in BC during naive, learning, and expert phases (Figure 4a). BC activity was higher in CR-Hit compared to Hit-Hit especially during learning and expert phases. The grand average activity for all 25 cortical areas highlights history-dependent information that emerges during learning (Figure 4b). We note that other areas, for example, different association areas, also encoded trial history information especially during learning and expert phases. In addition, we present activity difference maps between CR-Hit and Hit-Hit conditions during the stim period (Figure 4—figure supplement 1a). These maps clearly show the highest trial history information (i.e., difference in activity) in BC. Taken together, these results indicate that BC encodes trial history information that emerges during the stim period and just after learning. These results gave us the motivation to examine history-dependent information at time periods before texture touch.

Figure 4 with 1 supplement see all

Download asset Open asset

Trial history in RL before sensation

We next focused our analysis on the pre period, just before texture touch (−1 to −0.6 s before texture stop). Mean activity maps during the pre period highlight activity in association area RL that is present for both CR-Hit and Hit-Hit pairs during the naive, learning, and expert phases (Figure 5a; Gilad and Helmchen, 2020). RL pre period activity is higher in CR-Hit compared to Hit-Hit mostly during learning and expert phases. In addition, higher RL activity in CR-Hit pair starts even before the pre period, indicating that trial history is not directly related to the current stimulus (Figure 5b). The grand average of all 25 cortical areas, highlights the emergence of trial history during learning, especially in RL, but also in other association and sensory areas (Figure 5c). In addition, we present activity difference maps between CR-Hit and Hit-Hit conditions during the pre period (Figure 4—figure supplement 1b). These maps localize trial history information to RL which also spreads to other adjacent association areas. Moreover, activity patches slightly vary across the different mice which may affect the grand average (averaged across mice) of each area.

Figure 5 with 1 supplement see all

Download asset Open asset

RL activity was significantly higher in CR-Hit compared to Hit-Hit trials in the pre period during the expert phase (Figure 5—figure supplement 1; signed-rank test, p < 0.05, similar trend for the learning phase but insignificant; not significant for the naive phase). In addition, a statistical comparison between CR-Hit and Hit-Hit responses in RL within each mouse separately maintained significance for expert phase (5/7 mice; Mann–Whitney U-test p < 0.05). Body movements and whisker parameters did not significantly differ between CR-Hit and Hit-Hit conditions during the pre period (Similar to the stim period. Across and within mice. p > 0.05; Mann–Whitney U-test). The onset of history modulation within the trial structure (as in Figure 3d) was earlier in RL compared to BC in both learning (−0.15 ± 0.85 and 0.05 ± 0.86 s, median ± SD in RL and BC, respectively) and expert phases (−0.75 ± 0.53 and −0.1 ± 0.74 s, median ± SD in RL and BC, respectively) but not significantly different (p > 0.05, signed-rank test). The onset for the history modulation with relation to the learning profile in RL (similar to Figure 3h; During the pre period) was also earlier than BC, but not significantly different (200 ± 431 trials after crossing threshold compared to 500 ± 221 in BC; median ± SD, p > 0.05 singed rank test). Taken together, these results indicate that as mice gain expertise, prior to the sensation period, RL encodes history information, which may be later projected down onto BC where it is integrated with information of the current incoming texture.

Past versus present discrimination power in BC and RL

How well can BC and RL activity discriminate at the single trial level past information compared to the information of the current stimulus? To do this, we computed the receiver operating characteristics (ROC) analysis between specific trial types (Gilad et al., 2020; Gilad and Helmchen, 2020), along with the area under the curve (AUC) quantifying the discrimination power at the single trial level (Materials and methods). We calculated the AUC between two types of trials (Figure 6a): (1) Activity between CR-Hit and Hit-Hit pairs based on the activity during the Hit trial. This is defined as History AUC since only the previous trial is different. (2) Activity between the current Hit and CR trials. This is defined as the current AUC because the current trial types are different (both in terms of stimulus type and action). Both history and current AUCs are calculated for BC and RL for each time frame along the trial structure and for naive, learning, and expert phases. Intuitively, one would assume that the current AUC will display higher discrimination power compared to the history AUC because the latter AUC measure compares the same previous trial type which should be harder to discriminate. Interestingly, during the expert phase, history AUC in both BC and RL has a discrimination power in the stim period that is not significantly different than that of the current AUC (Figure 6b, c; p > 0.05; singed rank test). In other words, we found that BC and RL discriminate past stimuli just as well as the current stimuli. In addition, during the learning phase, RL and to some extent BC, display a significantly higher history AUC compared to the current AUC, specifically in the pre period (Figure 6d, e; p < 0.05; singed rank test). This indicates that history information is discriminative at the single trial level before stimulus onset. Taken together, we find that BC and RL can encode the past just as well as the present.

Figure 6

Download asset Open asset

We next calculated the history AUC for each pixel during either the pre or stim period (during expert phase). The history AUC maps during the pre period display AUC values around the RL areas (Figure 6f). In contrast, the history AUC maps during the stim period display AUC values mostly in BC (Figure 6g). Quantified across 25 areas and averaged across mice, RL displays the highest history AUC during the pre period, whereas BC displays the highest history AUC values during the stim period (Figure 6h, i). We note that additional cortical areas such as other association areas also display high history AUC values. In summary, we find that trial history emerges in RL before the texture arrives and then shifts to BC during stimulus presentation.

The data and custom code that support the findings of this study are publicly available at https://doi.org/10.17605/OSF.IO/HKVC5.

1. 1. Marmor O

   (2022) Open Science Framework

   HistoryLearning.

   https://doi.org/10.17605/OSF.IO/HKVC5

1. 1. Akrami A
   2. Kopec CD
   3. Diamond ME
   4. Brody CD

   (2018) Posterior parietal cortex represents sensory history and mediates its effects on behaviour

   Nature 554:368–372.

   https://doi.org/10.1038/nature25510

   • PubMed
   • Google Scholar
2. 1. Banerjee A
   2. Parente G
   3. Teutsch J
   4. Lewis C
   5. Voigt FF
   6. Helmchen F

   (2020) Value-guided remapping of sensory cortex by lateral orbitofrontal cortex

   Nature 585:245–250.

   https://doi.org/10.1038/s41586-020-2704-z

   • PubMed
   • Google Scholar
3. 1. Blake DT
   2. Strata F
   3. Churchland AK
   4. Merzenich MM

   (2002) Neural correlates of instrumental learning in primary auditory cortex

   PNAS 99:10114–10119.

   https://doi.org/10.1073/pnas.092278099

   • PubMed
   • Google Scholar
4. 1. Chen JL
   2. Carta S
   3. Soldado-Magraner J
   4. Schneider BL
   5. Helmchen F

   (2013) Behaviour-dependent recruitment of long-range projection neurons in somatosensory cortex

   Nature 499:336–340.

   https://doi.org/10.1038/nature12236

   • PubMed
   • Google Scholar
5. 1. Chen JL
   2. Margolis DJ
   3. Stankov A
   4. Sumanovski LT
   5. Schneider BL
   6. Helmchen F

   (2015) Pathway-specific reorganization of projection neurons in somatosensory cortex during learning

   Nature Neuroscience 18:1101–1108.

   https://doi.org/10.1038/nn.4046

   • PubMed
   • Google Scholar
6. 1. Chéreau R
   2. Bawa T
   3. Fodoulian L
   4. Carleton A
   5. Pagès S
   6. Holtmaat A

   (2020) Dynamic perceptual feature selectivity in primary somatosensory cortex upon reversal learning

   Nature Communications 11:3245.

   https://doi.org/10.1038/s41467-020-17005-x

   • PubMed
   • Google Scholar
7. 1. Driscoll LN
   2. Pettit NL
   3. Minderer M
   4. Chettih SN
   5. Harvey CD

   (2017) Dynamic reorganization of neuronal activity patterns in parietal cortex

   Cell 170:986–999.

   https://doi.org/10.1016/j.cell.2017.07.021

   • PubMed
   • Google Scholar
8. 1. Estebanez L
   2. El Boustani S
   3. Destexhe A
   4. Shulz DE

   (2012) Correlated input reveals coexisting coding schemes in a sensory cortex

   Nature Neuroscience 15:1691–1699.

   https://doi.org/10.1038/nn.3258

   • PubMed
   • Google Scholar
9. 1. Fu Y
   2. Kaneko M
   3. Tang Y
   4. Alvarez-Buylla A
   5. Stryker MP

   (2015) A cortical disinhibitory circuit for enhancing adult plasticity

   eLife 4:e05558.

   https://doi.org/10.7554/eLife.05558

   • PubMed
   • Google Scholar
10. 1. Gallero-Salas Y
    2. Han S
    3. Sych Y
    4. Voigt FF
    5. Laurenczy B
    6. Gilad A
    7. Helmchen F

    (2021) Sensory and behavioral components of neocortical signal flow in discrimination tasks with short-term memory

    Neuron 109:135–148.

    https://doi.org/10.1016/j.neuron.2020.10.017

    • PubMed
    • Google Scholar
11. 1. Garion L
    2. Dubin U
    3. Rubin Y
    4. Khateb M
    5. Schiller Y
    6. Azouz R
    7. Schiller J

    (2014) Texture coarseness responsive neurons and their mapping in layer 2-3 of the rat barrel cortex in vivo

    eLife 3:e03405.

    https://doi.org/10.7554/eLife.03405

    • PubMed
    • Google Scholar
12. 1. Garner AR
    2. Rowland DC
    3. Hwang SY
    4. Baumgaertel K
    5. Roth BL
    6. Kentros C
    7. Mayford M

    (2012) Generation of a synthetic memory trace

    Science 335:1513–1516.

    https://doi.org/10.1126/science.1214985

    • PubMed
    • Google Scholar
13. 1. Garrett M
    2. Manavi S
    3. Roll K
    4. Ollerenshaw DR
    5. Groblewski PA
    6. Ponvert ND
    7. Kiggins JT
    8. Casal L
    9. Mace K
    10. Williford A
    11. Leon A
    12. Jia X
    13. Ledochowitsch P
    14. Buice MA
    15. Wakeman W
    16. Mihalas S
    17. Olsen SR

    (2020) Experience shapes activity dynamics and stimulus coding of VIP inhibitory cells

    eLife 9:e50340.

    https://doi.org/10.7554/eLife.50340

    • PubMed
    • Google Scholar
14. Software

    1. Gilad A

    (2016) Behavioral go-Nogo program

    OSF.

    http://osf.io/hkvc5
15. 1. Gilad A
    2. Gallero-Salas Y
    3. Groos D
    4. Helmchen F

    (2018) Behavioral strategy determines frontal or posterior location of short-term memory in neocortex

    Neuron 99:814–828.

    https://doi.org/10.1016/j.neuron.2018.07.029

    • PubMed
    • Google Scholar
16. 1. Gilad A
    2. Helmchen F

    (2020) Spatiotemporal refinement of signal flow through association cortex during learning

    Nature Communications 11:1744.

    https://doi.org/10.1038/s41467-020-15534-z

    • PubMed
    • Google Scholar
17. 1. Gilad A
    2. Maor I
    3. Mizrahi A

    (2020) Learning-related population dynamics in the auditory thalamus

    eLife 9:e56307.

    https://doi.org/10.7554/eLife.56307

    • PubMed
    • Google Scholar
18. 1. Gossen M
    2. Bujard H

    (1992) Tight control of gene expression in mammalian cells by tetracycline-responsive promoters

    PNAS 89:5547–5551.

    https://doi.org/10.1073/pnas.89.12.5547

    • PubMed
    • Google Scholar
19. 1. Guo W
    2. Yao D
    3. Jiang J
    4. Su Q
    5. Zhang Z
    6. Zhang J
    7. Yu L
    8. Xiao C

    (2014) Abnormal default-mode network homogeneity in first-episode, drug-naive schizophrenia at rest

    Progress in Neuro-Psychopharmacology and Biological Psychiatry 49:16–20.

    https://doi.org/10.1016/j.pnpbp.2013.10.021

    • Google Scholar
20. 1. Harvey CD
    2. Coen P
    3. Tank DW

    (2012) Choice-specific sequences in parietal cortex during a virtual-navigation decision task

    Nature 484:62–68.

    https://doi.org/10.1038/nature10918

    • PubMed
    • Google Scholar
21. 1. Hattori R
    2. Danskin B
    3. Babic Z
    4. Mlynaryk N
    5. Komiyama T

    (2019) Area-specificity and plasticity of history-dependent value coding during learning

    Cell 177:1858–1872.

    https://doi.org/10.1016/j.cell.2019.04.027

    • PubMed
    • Google Scholar
22. 1. Hovde K
    2. Gianatti M
    3. Witter MP
    4. Whitlock JR

    (2019) Architecture and organization of mouse posterior parietal cortex relative to extrastriate areas

    The European Journal of Neuroscience 49:1313–1329.

    https://doi.org/10.1111/ejn.14280

    • PubMed
    • Google Scholar
23. 1. Hwang EJ
    2. Dahlen JE
    3. Mukundan M
    4. Komiyama T

    (2017) History-based action selection bias in posterior parietal cortex

    Nature Communications 8:1242.

    https://doi.org/10.1038/s41467-017-01356-z

    • PubMed
    • Google Scholar
24. 1. Johnson CM
    2. Peckler H
    3. Tai LH
    4. Wilbrecht L

    (2016) Rule learning enhances structural plasticity of long-range axons in frontal cortex

    Nature Communications 7:10785.

    https://doi.org/10.1038/ncomms10785

    • PubMed
    • Google Scholar
25. 1. Jurjut O
    2. Georgieva P
    3. Busse L
    4. Katzner S

    (2017) Learning enhances sensory processing in mouse v1 before improving behavior

    The Journal of Neuroscience 37:6460–6474.

    https://doi.org/10.1523/JNEUROSCI.3485-16.2017

    • PubMed
    • Google Scholar
26. 1. Kawai T
    2. Yamada H
    3. Sato N
    4. Takada M
    5. Matsumoto M

    (2015) Roles of the lateral habenula and anterior cingulate cortex in negative outcome monitoring and behavioral adjustment in nonhuman primates

    Neuron 88:792–804.

    https://doi.org/10.1016/j.neuron.2015.09.030

    • PubMed
    • Google Scholar
27. 1. Khodagholy D
    2. Gelinas JN
    3. Buzsáki G

    (2017) Learning-enhanced coupling between ripple oscillations in association cortices and hippocampus

    Science 358:369–372.

    https://doi.org/10.1126/science.aan6203

    • PubMed
    • Google Scholar
28. 1. Knutsen PM
    2. Derdikman D
    3. Ahissar E

    (2005) Tracking whisker and head movements in unrestrained behaving rodents

    Journal of Neurophysiology 93:2294–2301.

    https://doi.org/10.1152/jn.00718.2004

    • PubMed
    • Google Scholar
29. 1. Komiyama T
    2. Sato TR
    3. O’Connor DH
    4. Zhang Y-X
    5. Huber D
    6. Hooks BM
    7. Gabitto M
    8. Svoboda K

    (2010) Learning-related fine-scale specificity imaged in motor cortex circuits of behaving mice

    Nature 464:1182–1186.

    https://doi.org/10.1038/nature08897

    • PubMed
    • Google Scholar
30. 1. Lacefield CO
    2. Pnevmatikakis EA
    3. Paninski L
    4. Bruno RM

    (2019) Reinforcement learning recruits somata and apical dendrites across layers of primary sensory cortex

    Cell Reports 26:2000–2008.

    https://doi.org/10.1016/j.celrep.2019.01.093

    • PubMed
    • Google Scholar
31. 1. Le Merre P
    2. Esmaeili V
    3. Charrière E
    4. Galan K
    5. Salin PA
    6. Petersen CCH
    7. Crochet S

    (2018) Reward-based learning drives rapid sensory signals in medial prefrontal cortex and dorsal hippocampus necessary for goal-directed behavior

    Neuron 97:83–91.

    https://doi.org/10.1016/j.neuron.2017.11.031

    • PubMed
    • Google Scholar
32. 1. Li W
    2. Piëch V
    3. Gilbert CD

    (2008) Learning to link visual contours

    Neuron 57:442–451.

    https://doi.org/10.1016/j.neuron.2007.12.011

    • PubMed
    • Google Scholar
33. 1. Lyamzin D
    2. Benucci A

    (2019) The mouse posterior parietal cortex: Anatomy and functions

    Neuroscience Research 140:14–22.

    https://doi.org/10.1016/j.neures.2018.10.008

    • PubMed
    • Google Scholar
34. 1. Makino H
    2. Komiyama T

    (2015) Learning enhances the relative impact of top-down processing in the visual cortex

    Nature Neuroscience 18:1116–1122.

    https://doi.org/10.1038/nn.4061

    • Google Scholar
35. 1. Mathis A
    2. Mamidanna P
    3. Cury KM
    4. Abe T
    5. Murthy VN
    6. Mathis MW
    7. Bethge M

    (2018) DeepLabCut: markerless pose estimation of user-defined body parts with deep learning

    Nature Neuroscience 21:1281–1289.

    https://doi.org/10.1038/s41593-018-0209-y

    • PubMed
    • Google Scholar
36. Preprint

    1. Mohan H
    2. An X
    3. Kondo H
    4. Zhao S
    5. Matho KS
    6. Musall S
    7. Mitra P
    8. Huang ZJ

    (2022) Cortical Glutamatergic Projection Neuron Types Contribute to Distinct Functional Subnetworks

    bioRxiv.

    https://doi.org/10.1101/2021.12.30.474537

    • Google Scholar
37. 1. Morcos AS
    2. Harvey CD

    (2016) History-dependent variability in population dynamics during evidence accumulation in cortex

    Nature Neuroscience 19:1672–1681.

    https://doi.org/10.1038/nn.4403

    • Google Scholar
38. 1. Musall S
    2. Sun XR
    3. Mohan H
    4. An X
    5. Gluf S
    6. Li SJ
    7. Drewes R
    8. Cravo E
    9. Lenzi I
    10. Yin C
    11. Kampa BM
    12. Churchland AK

    (2023) Pyramidal cell types drive functionally distinct cortical activity patterns during decision-making

    Nature Neuroscience 26:495–505.

    https://doi.org/10.1038/s41593-022-01245-9

    • PubMed
    • Google Scholar
39. 1. Oh SW
    2. Harris JA
    3. Ng L
    4. Winslow B
    5. Cain N
    6. Mihalas S
    7. Wang Q
    8. Lau C
    9. Kuan L
    10. Henry AM
    11. Mortrud MT
    12. Ouellette B
    13. Nguyen TN
    14. Sorensen SA
    15. Slaughterbeck CR
    16. Wakeman W
    17. Li Y
    18. Feng D
    19. Ho A
    20. Nicholas E
    21. Hirokawa KE
    22. Bohn P
    23. Joines KM
    24. Peng H
    25. Hawrylycz MJ
    26. Phillips JW
    27. Hohmann JG
    28. Wohnoutka P
    29. Gerfen CR
    30. Koch C
    31. Bernard A
    32. Dang C
    33. Jones AR
    34. Zeng H

    (2014) A mesoscale connectome of the mouse brain

    Nature 508:207–214.

    https://doi.org/10.1038/nature13186

    • PubMed
    • Google Scholar
40. 1. Pasupathy A
    2. Miller EK

    (2005) Different time courses of learning-related activity in the prefrontal cortex and striatum

    Nature 433:873–876.

    https://doi.org/10.1038/nature03287

    • PubMed
    • Google Scholar
41. 1. Petreanu L
    2. Gutnisky DA
    3. Huber D
    4. Xu N
    5. O’Connor DH
    6. Tian L
    7. Looger L
    8. Svoboda K

    (2012) Activity in motor-sensory projections reveals distributed coding in somatosensation

    Nature 489:299–303.

    https://doi.org/10.1038/nature11321

    • PubMed
    • Google Scholar
42. 1. Pfeffer CK
    2. Xue M
    3. He M
    4. Huang ZJ
    5. Scanziani M

    (2013) Inhibition of inhibition in visual cortex: the logic of connections between molecularly distinct interneurons

    Nature Neuroscience 16:1068–1076.

    https://doi.org/10.1038/nn.3446

    • PubMed
    • Google Scholar
43. 1. Pho GN
    2. Goard MJ
    3. Woodson J
    4. Crawford B
    5. Sur M

    (2018) Task-dependent representations of stimulus and choice in mouse parietal cortex

    Nature Communications 9:2596.

    https://doi.org/10.1038/s41467-018-05012-y

    • PubMed
    • Google Scholar
44. 1. Poort J
    2. Khan AG
    3. Pachitariu M
    4. Nemri A
    5. Orsolic I
    6. Krupic J
    7. Bauza M
    8. Sahani M
    9. Keller GB
    10. Mrsic-Flogel TD
    11. Hofer SB

    (2015) Learning enhances sensory and multiple non-sensory representations in primary visual cortex

    Neuron 86:1478–1490.

    https://doi.org/10.1016/j.neuron.2015.05.037

    • PubMed
    • Google Scholar
45. 1. Rodgers CC
    2. Nogueira R
    3. Pil BC
    4. Greeman EA
    5. Park JM
    6. Hong YK
    7. Fusi S
    8. Bruno RM

    (2021) Sensorimotor strategies and neuronal representations for shape discrimination

    Neuron 109:2308–2325.

    https://doi.org/10.1016/j.neuron.2021.05.019

    • PubMed
    • Google Scholar
46. 1. Roelfsema PR
    2. Holtmaat A

    (2018) Control of synaptic plasticity in deep cortical networks

    Nature Reviews. Neuroscience 19:166–180.

    https://doi.org/10.1038/nrn.2018.6

    • PubMed
    • Google Scholar
47. 1. Safaai H
    2. von Heimendahl M
    3. Sorando JM
    4. Diamond ME
    5. Maravall M

    (2013) Coordinated population activity underlying texture discrimination in rat barrel cortex

    The Journal of Neuroscience 33:5843–5855.

    https://doi.org/10.1523/JNEUROSCI.3486-12.2013

    • PubMed
    • Google Scholar
48. 1. Save E
    2. Poucet B

    (2009) Role of the parietal cortex in long-term representation of spatial information in the rat

    Neurobiology of Learning and Memory 91:172–178.

    https://doi.org/10.1016/j.nlm.2008.08.005

    • PubMed
    • Google Scholar
49. 1. Scott BB
    2. Constantinople CM
    3. Akrami A
    4. Hanks TD
    5. Brody CD
    6. Tank DW

    (2017) Fronto-parietal cortical circuits encode accumulated evidence with a diversity of timescales

    Neuron 95:385–398.

    https://doi.org/10.1016/j.neuron.2017.06.013

    • PubMed
    • Google Scholar
50. 1. Silasi G
    2. Xiao D
    3. Vanni MP
    4. Chen ACN
    5. Murphy TH

    (2016) Intact skull chronic windows for mesoscopic wide-field imaging in awake mice

    Journal of Neuroscience Methods 267:141–149.

    https://doi.org/10.1016/j.jneumeth.2016.04.012

    • PubMed
    • Google Scholar
51. 1. Sul JH
    2. Kim H
    3. Huh N
    4. Lee D
    5. Jung MW

    (2010) Distinct roles of rodent orbitofrontal and medial prefrontal cortex in decision making

    Neuron 66:449–460.

    https://doi.org/10.1016/j.neuron.2010.03.033

    • PubMed
    • Google Scholar
52. 1. Suzuki A
    2. Kosugi S
    3. Murayama E
    4. Sasakawa E
    5. Ohkawa N
    6. Konno A
    7. Hirai H
    8. Inokuchi K

    (2022) A cortical cell ensemble in the posterior parietal cortex controls past experience-dependent memory updating

    Nature Communications 13:41.

    https://doi.org/10.1038/s41467-021-27763-x

    • PubMed
    • Google Scholar
53. 1. Tsutsui KI
    2. Grabenhorst F
    3. Kobayashi S
    4. Schultz W

    (2016) A dynamic code for economic object valuation in prefrontal cortex neurons

    Nature Communications 7:12554.

    https://doi.org/10.1038/ncomms12554

    • PubMed
    • Google Scholar
54. 1. Vann SD
    2. Aggleton JP
    3. Maguire EA

    (2009) What does the retrosplenial cortex do?

    Nature Reviews. Neuroscience 10:792–802.

    https://doi.org/10.1038/nrn2733

    • PubMed
    • Google Scholar
55. 1. Vanni MP
    2. Murphy TH

    (2014) Mesoscale transcranial spontaneous activity mapping in GCaMP3 transgenic mice reveals extensive reciprocal connections between areas of somatomotor cortex

    The Journal of Neuroscience 34:15931–15946.

    https://doi.org/10.1523/JNEUROSCI.1818-14.2014

    • PubMed
    • Google Scholar
56. 1. Vecchia D
    2. Beltramo R
    3. Vallone F
    4. Chéreau R
    5. Forli A
    6. Molano-Mazón M
    7. Bawa T
    8. Binini N
    9. Moretti C
    10. Holtmaat A
    11. Panzeri S
    12. Fellin T

    (2020) Temporal sharpening of sensory responses by layer v in the mouse primary somatosensory cortex

    Current Biology 30:1589–1599.

    https://doi.org/10.1016/j.cub.2020.02.004

    • PubMed
    • Google Scholar
57. 1. Whitlock JR
    2. Sutherland RJ
    3. Witter MP
    4. Moser M-B
    5. Moser EI

    (2008) Navigating from hippocampus to parietal cortex

    PNAS 105:14755–14762.

    https://doi.org/10.1073/pnas.0804216105

    • Google Scholar
58. 1. Williams LE
    2. Holtmaat A

    (2019) Higher-order thalamocortical inputs gate synaptic long-term potentiation via disinhibition

    Neuron 101:91–102.

    https://doi.org/10.1016/j.neuron.2018.10.049

    • PubMed
    • Google Scholar
59. 1. Yan Y
    2. Rasch MJ
    3. Chen M
    4. Xiang X
    5. Huang M
    6. Wu S
    7. Li W

    (2014) Perceptual training continuously refines neuronal population codes in primary visual cortex

    Nature Neuroscience 17:1380–1387.

    https://doi.org/10.1038/nn.3805

    • PubMed
    • Google Scholar
60. 1. Zuo Y
    2. Diamond ME

    (2019) Texture identification by bounded integration of sensory cortical signals

    Current Biology 29:1425–1435.

    https://doi.org/10.1016/j.cub.2019.03.017

    • PubMed
    • Google Scholar

Author details

1. Odeya Marmor

   Department of Medical Neurobiology, Faculty of Medicine, The Institute for Medical Research Israel-Canada (IMRIC), The Hebrew University of Jerusalem, Jerusalem, Israel

   Contribution

   Conceptualization, Software, Formal analysis, Visualization, Writing - original draft

   For correspondence

   odeya.marmor@gmail.com

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-0290-5888

2. Yael Pollak

   Department of Medical Neurobiology, Faculty of Medicine, The Institute for Medical Research Israel-Canada (IMRIC), The Hebrew University of Jerusalem, Jerusalem, Israel

   Contribution

   Data curation, conducted additional control experiments

   Competing interests

   No competing interests declared

3. Chen Doron

   Department of Medical Neurobiology, Faculty of Medicine, The Institute for Medical Research Israel-Canada (IMRIC), The Hebrew University of Jerusalem, Jerusalem, Israel

   Contribution

   Data curation, conducted additional control experiments

   Competing interests

   No competing interests declared

4. Fritjof Helmchen

   1. Brain Research Institute, University of Zurich, Zurich, Switzerland
   2. Neuroscience Center Zurich, Zurich, Switzerland

   Contribution

   Resources, Funding acquisition, Methodology, Writing - review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-8867-9569

5. Ariel Gilad

   Department of Medical Neurobiology, Faculty of Medicine, The Institute for Medical Research Israel-Canada (IMRIC), The Hebrew University of Jerusalem, Jerusalem, Israel

   Contribution

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

   For correspondence

   ariel.gilad@mail.huji.ac.il

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-8802-8611

• 1,268

  views

• 105

  downloads

• 4

  citations

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