In the primary visual cortex (V1), layer 2 and 3 (L2/3) pyramidal neurons conform to a stereotypical morphology characterized by separate apical and basal dendritic arbors. The apical tree consists of a thick apical trunk that extends into L1 and splits into an apical tuft. The basal tree consists of numerous dendritic segments, the majority of which sprout directly from the base of the soma, indicating that they can directly influence somatic output (Spruston, 2008). This compartmentalization of V1 neurons is also evident in the wiring diagram describing each tree: layer 4 (L4) and L2/3 pyramidal neurons synapse with basal dendrites of L2/3 pyramidal neurons, providing them with feedforward input (Coogan and Burkhalter, 1990; Larkum, 2013). The apical dendrites of the L2/3 pyramidal neurons instead receive feedback input from higher order areas of the cortex such as the secondary and tertiary visual cortices, lateromedial area and prefrontal cortex (Coogan and Burkhalter, 1990; Larkum, 2013), as well as orientation-tuned thalamocortical input (Chen et al., 2013; Cruz-Martín et al., 2014; Jia et al., 2010; Roth et al., 2016). These anatomical and connectivity features of the apical and basal trees, complemented by their distinct biophysical properties (Cho et al., 2008), shape both the local dendritic processing and the integration of the two input streams at the soma.

Few studies have undertaken the technically difficult task of measuring single spine tuning properties in the mouse visual cortex (Chen et al., 2013; Iacaruso et al., 2017; Jia et al., 2010). These studies have shown that single spines of L2/3 pyramidal neurons exhibit orientation selectivity, with synaptic orientation preferences varying even on the same branch (Iacaruso et al., 2017; Jia et al., 2010). These findings raise the question of how dendritic integration of sparse synaptic inputs along the apical and basal trees shape orientation tuning in L2/3 V1 pyramidal neurons.

The effect of synaptic input on neuronal output depends on the biophysical properties of dendritic branches. Older modeling work has established that blockage of dendritic sodium and NMDA activity leads to impaired or abolished orientation selectivity at the level of the soma (Archie and Mel, 2000; Mel et al., 1998). It is also known that individual dendritic segments of L2/3 neurons generate nonlinear regenerative responses, termed dendritic spikes, independently of the soma, making their exact contribution to somatic output unclear (Palmer et al., 2014; Smith et al., 2013). This could indicate that multiple dendritic segments need to be activated in order to generate a spike at the soma. On the other hand, it was recently proposed that strong, sparse inputs to the dendrites are sufficient to drive somatic output (Goetz et al., 2021). Interestingly, although dendritic spikes in the apical tuft of L2/3 V1 pyramidal neurons influence orientation selectivity (Goetz et al., 2021; Smith et al., 2013), this orientation selectivity is robust to ablation of the apical tree (Park et al., 2019). Altogether, the above studies highlight the need to scrutinize the functional effect that synaptic inputs along the basal and apical dendritic arbors have on the soma, building towards a realistic theory of visual processing.

In this work, we used a detailed biophysical model of a L2/3 V1 pyramidal cell (Park et al., 2019), to tease apart the relative contributions of apical vs. basal dendritic trees in somatic orientation tuning. We start by examining the threshold and strength of dendritic non-linearities, which are found to vary along the apical and basal dendrites. By comparing model output to in vivo two-photon calcium fluorescence imaging data, we find that dendritic and somatic spontaneous activity patterns in the model adequately reproduce experimental data. We then simulate orientation tuning in our model and evaluate its responses to specific combinations of dendritic disparity and synaptic distributions. We find that the tuned output of the model does not always match the linear summation of tuned input, instead deviating towards orientation preferences more closely matching either the apical or basal mean orientation preference.

We continue by evaluating the dependence of somatic spikes on dendritic voltage-gated sodium channel conductances by performing just-in-time interventions that nullify sodium conductance on either the apical or basal tree prior to a somatic spike. We find that the majority of stimulus-driven somatic spiking activity is dependent on sodium spikes generated in the apical tree, suggesting that apical and basal trees contribute via different mechanisms to tuned somatic output. We investigate this further by examining the role of both specific ionic (sodium channels) and synaptic (AMPA/NMDA) mechanisms in shaping orientation tuning. Results indicate that apical sodium channels are critical for proper neuronal tuning, while basal sodium channels do not play as significant a role. Conversely, AMPA and NMDA activity of both the apical and basal trees is required for orientation tuning at the level of the soma. A potential reason for this is that basal dendritic branches feature higher sodium spiking thresholds for similar electrotonic constant-normalized length values (i.e. ‘electrotonic length’) compared to apical branches, reducing their propensity to generate sodium spikes compared to apical dendrites.

Overall, our simulations suggest a synergistic effect between apical and basal trees, as somatic spikes are only reliably produced when apical dendritic sodium spikes coincide with basal synaptically driven depolarizations and/or spikes. In this manner, apical and basal dendrites contribute to neuronal function using different mechanisms, ionic or synaptic, which together drive action potential generation in this neuronal type.

In this work, we used a detailed computational model to delineate the dendritic constituents of neuronal computation in L2/3 V1 pyramidal cells. We found that the model exhibits a wide range of non-linear behavior at the dendritic segment level (Figure 1) and can mimic the activity patterns of neurons under spontaneous conditions in vivo (Figure 2—figure supplement 1). The model neuron exhibits robust orientation tuning in response to tuned input (Figure 2A), and under conditions of dendritic disparity, the model still remains robustly tuned (OSI <0.2 and tuning width >80°) for non-extreme dendritic disparity conditions (disparity ≤50°). Importantly, the neuronal orientation preference deviates from the expected one (Figure 2D–F), indicating that apical and basal inputs interact in non-linear ways. This bias is dependent on the distribution of synaptic inputs along the two trees (Figure 2D–F), it follows the preference of basal tree inputs under physiological conditions (Figure 2D) and is in line with prior work, whereby apical tree ablation had no impact on somatic orientation preference (Figure 2—figure supplement 5; Park et al., 2019). Remarkably, while the above would seem to suggest a minor role for apical inputs in the orientation preference of our model cell (Figure 2D), we found that the orientation tuning of the neuron is greatly affected by sodium conductances in the apical – but not the basal – tree (Figure 4). Interestingly, while basal sodium channel conductances are not critically important (Figure 4C), AMPA and NMDA conductances are necessary in both dendritic trees for orientation tuning (Figure 4A, B). This indicates that depolarizations from the basal tree play a key role in shaping neuronal output, even when basal sodium spikes do not. As such, neuronal output is a synergistic phenomenon mostly between apical sodium spikes and basal depolarizations or sodium spikes. Our simulations suggest that a form of intra-tree dendritic cooperativity – a synergistic effect between apical and basal dendritic segments that exhibit activity in relative synchrony – might be present in these neurons. This is supported by the existence of extinction-prone ‘cooperative’ somatic spikes that are lost when either of their dendritic components (apical or basal, be that depolarization or spiking) is lost (Figure 3). Additional evidence is provided by the experiments in which removal or suppression (decrease in conductance) of basal stimulus-driven synapses (i.e. only background-driven synapses are fully active) significantly decreases the quality of orientation tuning (i.e. OSI values) in the model neuron, or abolishes tuning altogether (Figure 4A, B).

A potential explanation for the differential contributions of apical and basal dendrites lies with their morphological and electrophysiological properties. Despite featuring a uniform sodium conductance throughout both dendritic trees, basal dendrites have higher sodium spiking thresholds than apical dendrites with similar electrotonic length values. Moreover, apical dendrites have highly variable spiking thresholds (Figure 5D). This is likely because the soma acts as a current sink, effectively increasing the sodium spiking threshold of all soma-proximal branches (mostly basal dendrites). Meanwhile, the majority of apical dendrites feature much better electrical isolation from the soma due to distance and branching, and are as such exempt from such current sink effects, instead only suffering a loss of current from other connected apical dendritic branches. As such, they exhibit a much greater range of possible thresholds compared to basal dendrites of similar volume (Figure 5D), while simultaneously featuring greater signal attenuation (Figure 5—figure supplement 1C, E). Put together, these results indicate that apical dendrites overcome their signal attenuation deficit (compared to basal dendrites) by having a greater propensity for dendritic sodium spike generation for different amounts of input. This suggests that sodium spikes are more difficult to induce in basal compared to apical dendrites, offering a potential explanation as to why apical sodium conductances contribute more to somatic output.

Taken together, our model predicts that somatic output in L2/3 V1 pyramidal neurons is likely to be determined through Bimodal Input Coincidence, a subcategory of coincidence detection (Figure 6). Specifically, we propose that the basal tree receives visual feedforward input, representing the information therein as a series of hyperpolarizations, depolarizations, and occasional dendritic spikes. Thus, visual input creates a basal ‘backdrop’ of depolarizations that represents visual information. At the same time, predictive and attentional signals from higher-order cortical areas reach the apical tree of the neuron, causing the generation of dendritic spikes that propagate to the soma and are temporally summed with any concurrent basal depolarization. In the event that visual input is non-existent, or of a non-preferred orientation, the depolarization is minimal to none. Thus, the apical dendritic spike will not be significantly augmented through summation and will fail to produce a somatic response in the majority of cases. If there is visual input, however, especially of an orientation matching the preferred orientation of the basal tree, the aforementioned ‘backdrop’ will include multiple subthreshold depolarizations, perhaps even dendritic spikes. The apical dendritic spikes are thus likely to be temporally summed with these depolarizations and generate a somatic spike. As such, even though most somatic spikes will be generated through an apical tree dendritic spike, the cases in which this is possible in the first place will be dictated by the backdrop of depolarizations (and/or spikes) provided by the basal tree.

Figure 6

Download asset Open asset

The findings presented here propose a more central role for the apical tree than previously suggested. Specifically, in Park et al., 2019, we found that orientation selectivity was not abolished by in vivo laser ablation of the apical tree in L2/3 mouse V1 pyramidal neurons, remaining essentially unchanged following a recovery period of ~1 day. Importantly, the model pr‘sented here reproduces the in vivo ablation findings of the aforementioned study (see Extended methods, Ablation protocol; Figure 2—figure supplement 5). What is seen as a contradiction is resolved by considering that in both Park et al., 2019 and the simulations presented here, apical dendrite ablation is assumed to be accompanied by homeostatic alterations that counteract the increase in input resistance from the loss of the apical tree. These homeostatic mechanisms keep the firing rate of the neuron the same pre- and post-ablation, as seen experimentally (Park et al., 2019). Similarly, the neuron can potentially compensate for the loss of input from one of its dendritic trees via homeostatic mechanisms that amplify the effect of synaptic inputs on the other tree, an approach followed by Mel et al., 1998 in their efforts to demonstrate that V1 L5 pyramidal neurons were capable of exhibiting orientation tuned responses through either apical or basal inputs. Thus, the contribution of the apical tree to somatic spiking in the intact neuron can be obviated by homeostatic mechanisms following apical tree ablation.

Our findings suggest that L2/3 V1 pyramidal cells may perform salient feature extraction from visual input compatible with predictive coding (Rao and Ballard, 1999). When top-down signals reflecting attentional or predictive processes reach the apical tree, they ‘highlight’ appropriate parts of the signal ‘landscape’ generated by the visual input reaching the basal trees through the generation of apical dendritic sodium spikes. This could result in a neuron that is activated only when specific features are present in its receptive field, facilitating the processing of salient stimuli. Such phenomena have been observed experimentally as a result of attention (Simons and Chabris, 1999).

However, this work also suffers from a number of limitations. Our findings are inherently bound to the particular characteristics of the simulated morphology. To ensure their generality, further exploration of different neuronal morphologies of L2/3 V1 neurons is required. However, given that our findings correlate with specific anatomical and electrophysiological features of the model neuron (Figure 5, Figure 5—figure supplement 1), it would be possible to infer whether the behavior we observe in our model could be present in other neuronal morphologies as well. In addition, while repeating our analysis using other morphologies is beyond the scope of this study (validating model function and performing the extensive parameter exploration reported here is highly resource-intensive), our code and all analysis scripts are made publicly available, allowing interested scientists to apply our approach to other morphologies. The sensitivity analysis that we added in Sensitivity analysis (Figure 4—figure supplement 1) also further establishes the robustness of our findings within biologically relevant ranges of ionic conductances. We have taken all these steps in order to more conclusively demonstrate that our findings are likely to generalize to other L2/3 V1 pyramidal neurons. Furthermore, our model is limited by the lack of data on dendritic features such as signal attenuation and distribution of ion channel conductances. However, this has been counteracted at least to some extent by the use of experimental data (e.g. see Figure 2—figure supplements 1–4) and the sensitivity analysis, which demonstrates that our conclusions are robust to biologically relevant variations of the key conductances (Figure 4—figure supplement 1). It should be noted, however, that there could be multiple ‘solutions’ to the problem of model validation, that is multiple different sets of values for the model parameters that yield the desired output. As such, we cannot be absolutely certain that our particular configuration of parameter values is correct, although this is an issue faced by all models in general, and not exclusive to this work.

Additionally, the single-neuron model sacrifices much of the complexity found in the actual L2/3 network, particularly certain input features. First, we do not simulate in vivo-like input patterns, due to a lack of data. Instead, we use randomly generated Poisson spike trains, which are widely used in the field. Also, in our model, inhibitory inputs are not patterned according to different interneuron subtypes and are background-driven, lacking tuning and featuring firing rates lower than stimulus-driven excitatory synapses. A similar complexity loss is incurred by using a relatively simple stimulation protocol, compared to the alternating stimulation of the ‘drifting gradient’ protocol that is more widely used. Additionally, the model does not account for retinotopically shifted inhibitory inputs (e.g. Rossi et al., 2020). Retinotopic shift in inhibitory inputs would not affect our results, since stimulation is implicitly assumed to only take place at the center of the receptive field of the model neuron, a fact which would render any retinotopically offset neurons quiescent. Finally, although inhibitory inputs could be biased towards one dendritic tree over the other, the overall effect of such an imbalance would be a change in the net amount of excitation received by each tree – a scenario which we already explore in part (Figure 2 and Figure 3), finding tuning and somatic spike generation to be generally unaffected. The omission of these features that were unlikely to influence the presence or absence of orientation tuning was performed in an attempt to reduce the complexity (and thus computational requirements) of the model.

In terms of questions that still need to be answered, the exact nature of the attentional and predictive signals received by the apical tree remains to be deciphered. Secondly, the formation of the visual ‘backdrop’ as a result of basal tree depolarization merits further study. Encoding of visual information using mostly subthreshold depolarizations in the presence of noise is an interesting problem, possibly resolved through the effects of intra-tree dendritic cooperativity, where apical spikes sharpen responses to stimulus-driven input rather than noise, improving the signal-to-noise ratio (Poleg-Polsky, 2019). In addition, a number of differences emerge when comparing the visual system across different model species. In rodents, orientation-selective cells akin to V1 simple cells have been discovered in the thalamus itself (Scholl et al., 2013), and direct thalamocortical projections to L1 of V1 have also been observed (Roth et al., 2016). Furthermore, rodent V1 organization follows a dispersed salt-and-pepper motif, unlike the highly structured organization typical of higher mammals like primates and cats (Ohki and Reid, 2007). These differences are important to take into account when trying to infer general rules of function from information derived from rules of information processing in different animal models. Finally, perhaps the most interesting unanswered question is whether these computations take place elsewhere in the cortex in addition to the visual system. Predictive coding as a means of stimulus compression has already been proposed as a way to simplify visual perception (Rao and Ballard, 1999), and this might also be the case for other sensory or cognitive tasks. Larger-scale models of the visual system, featuring realistic input structure and properties, will most likely be required to explore these phenomena in depth. Regardless, further investigation is required in order to unravel the Gordian knot that is visual perception.

The code used to create the figures of this manuscript is publicly available in GitHub: https://github.com/kepetousakis/petousakis_etal_2023 (copy archived at Petousakis, 2023). The code for the model neuron is available in ModelDB at https://modeldb.science/267501.

1. 1. Petousakis KE
   2. Park J
   3. Papoutsi A
   4. Smirnakis S
   5. Poirazi P

   (2023) ModelDB

   ID 267501. L2/3 V1 Pyramidal Cell model (modified Park et al., 2019; a/n: 231185) (Petousakis et al., 2023).

   https://modeldb.science/267501

1. 1. Adesnik H
   2. Bruns W
   3. Taniguchi H
   4. Huang ZJ
   5. Scanziani M

   (2012) A neural circuit for spatial summation in visual cortex

   Nature 490:226–231.

   https://doi.org/10.1038/nature11526

   • PubMed
   • Google Scholar
2. 1. Archie KA
   2. Mel BW

   (2000) A model for intradendritic computation of binocular disparity

   Nature Neuroscience 3:54–63.

   https://doi.org/10.1038/71125

   • PubMed
   • Google Scholar
3. 1. Behabadi BF
   2. Polsky A
   3. Jadi M
   4. Schiller J
   5. Mel BW

   (2012) Location-dependent excitatory synaptic interactions in pyramidal neuron dendrites

   PLOS Computational Biology 8:e1002599.

   https://doi.org/10.1371/journal.pcbi.1002599

   • PubMed
   • Google Scholar
4. 1. Chen TW
   2. Wardill TJ
   3. Sun Y
   4. Pulver SR
   5. Renninger SL
   6. Baohan A
   7. Schreiter ER
   8. Kerr RA
   9. Orger MB
   10. Jayaraman V
   11. Looger LL
   12. Svoboda K
   13. Kim DS

   (2013) Ultrasensitive fluorescent proteins for imaging neuronal activity

   Nature 499:295–300.

   https://doi.org/10.1038/nature12354

   • PubMed
   • Google Scholar
5. 1. Cho KH
   2. Jang HJ
   3. Lee EH
   4. Yoon SH
   5. Hahn SJ
   6. Jo YH
   7. Kim MS
   8. Rhie DJ

   (2008) Differential cholinergic modulation of Ca2+ transients evoked by backpropagating action potentials in apical and basal dendrites of cortical pyramidal neurons

   Journal of Neurophysiology 99:2833–2843.

   https://doi.org/10.1152/jn.00063.2008

   • PubMed
   • Google Scholar
6. 1. Cho KH
   2. Jang JH
   3. Jang HJ
   4. Kim MJ
   5. Yoon SH
   6. Fukuda T
   7. Tennigkeit F
   8. Singer W
   9. Rhie DJ

   (2010) Subtype-specific dendritic Ca(2+) dynamics of inhibitory interneurons in the rat visual cortex

   Journal of Neurophysiology 104:840–853.

   https://doi.org/10.1152/jn.00146.2010

   • PubMed
   • Google Scholar
7. 1. Coogan TA
   2. Burkhalter A

   (1990) Conserved patterns of cortico-cortical connections define areal hierarchy in rat visual cortex

   Experimental Brain Research 80:49–53.

   https://doi.org/10.1007/BF00228846

   • PubMed
   • Google Scholar
8. 1. Cruz-Martín A
   2. El-Danaf RN
   3. Osakada F
   4. Sriram B
   5. Dhande OS
   6. Nguyen PL
   7. Callaway EM
   8. Ghosh A
   9. Huberman AD

   (2014) A dedicated circuit links direction-selective retinal ganglion cells to the primary visual cortex

   Nature 507:358–361.

   https://doi.org/10.1038/nature12989

   • PubMed
   • Google Scholar
9. 1. Defelipe J
   2. Fariñas I

   (1992) The pyramidal neuron of the cerebral cortex: morphological and chemical characteristics of the synaptic inputs

   Progress in Neurobiology 39:563–607.

   https://doi.org/10.1016/0301-0082(92)90015-7

   • PubMed
   • Google Scholar
10. 1. Defelipe J
    2. Elston GN
    3. Fujita I
    4. Fuster J
    5. Harrison KH
    6. Hof PR
    7. Kawaguchi Y
    8. Martin KAC
    9. Rockland KS
    10. Thomson AM
    11. Wang SSH
    12. White EL
    13. Yuste R

    (2002) Neocortical circuits: evolutionary aspects and specificity versus non-specificity of synaptic connections

    Journal of Neurocytology 31:387–416.

    https://doi.org/10.1023/A:1024142513991

    • PubMed
    • Google Scholar
11. 1. Goetz L
    2. Roth A
    3. Häusser M

    (2021) Active dendrites enable strong but sparse inputs to determine orientation selectivity

    PNAS 118:e2017339118.

    https://doi.org/10.1073/pnas.2017339118

    • PubMed
    • Google Scholar
12. 1. Haider B
    2. Häusser M
    3. Carandini M

    (2013) Inhibition dominates sensory responses in the awake cortex

    Nature 493:97–100.

    https://doi.org/10.1038/nature11665

    • PubMed
    • Google Scholar
13. 1. Hines ML
    2. Carnevale NT

    (2001) NEURON: a tool for neuroscientists

    The Neuroscientist 7:123–135.

    https://doi.org/10.1177/107385840100700207

    • PubMed
    • Google Scholar
14. 1. Iacaruso MF
    2. Gasler IT
    3. Hofer SB

    (2017) Synaptic organization of visual space in primary visual cortex

    Nature 547:449–452.

    https://doi.org/10.1038/nature23019

    • PubMed
    • Google Scholar
15. 1. Jia H
    2. Rochefort NL
    3. Chen X
    4. Konnerth A

    (2010) Dendritic organization of sensory input to cortical neurons in vivo

    Nature 464:1307–1312.

    https://doi.org/10.1038/nature08947

    • PubMed
    • Google Scholar
16. 1. Ju N
    2. Li Y
    3. Liu F
    4. Jiang H
    5. Macknik SL
    6. Martinez-Conde S
    7. Tang S

    (2020) Spatiotemporal functional organization of excitatory synaptic inputs onto macaque V1 neurons

    Nature Communications 11:697.

    https://doi.org/10.1038/s41467-020-14501-y

    • PubMed
    • Google Scholar
17. 1. Keller AJ
    2. Roth MM
    3. Scanziani M

    (2020) Feedback generates a second receptive field in neurons of the visual cortex

    Nature 582:545–549.

    https://doi.org/10.1038/s41586-020-2319-4

    • PubMed
    • Google Scholar
18. 1. Ko H
    2. Hofer SB
    3. Pichler B
    4. Buchanan KA
    5. Sjöström PJ
    6. Mrsic-Flogel TD

    (2011) Functional specificity of local synaptic connections in neocortical networks

    Nature 473:87–91.

    https://doi.org/10.1038/nature09880

    • PubMed
    • Google Scholar
19. 1. Larkum M

    (2013) A cellular mechanism for cortical associations: an organizing principle for the cerebral cortex

    Trends in Neurosciences 36:141–151.

    https://doi.org/10.1016/j.tins.2012.11.006

    • PubMed
    • Google Scholar
20. 1. Mel BW
    2. Ruderman DL
    3. Archie KA

    (1998) Translation-invariant orientation tuning in visual “complex” cells could derive from intradendritic computations

    The Journal of Neuroscience 18:4325–4334.

    https://doi.org/10.1523/JNEUROSCI.18-11-04325.1998

    • PubMed
    • Google Scholar
21. 1. Mrsic-Flogel TD
    2. Hofer SB
    3. Ohki K
    4. Reid RC
    5. Bonhoeffer T
    6. Hübener M

    (2007) Homeostatic regulation of eye-specific responses in visual cortex during ocular dominance plasticity

    Neuron 54:961–972.

    https://doi.org/10.1016/j.neuron.2007.05.028

    • PubMed
    • Google Scholar
22. 1. Ohki K
    2. Reid RC

    (2007) Specificity and randomness in the visual cortex

    Current Opinion in Neurobiology 17:401–407.

    https://doi.org/10.1016/j.conb.2007.07.007

    • PubMed
    • Google Scholar
23. Preprint

    1. Pachitariu M
    2. Stringer C
    3. Dipoppa M
    4. Schröder S
    5. Rossi LF
    6. Dalgleish H
    7. Carandini M
    8. Harris KD

    (2016) Suite2p: Beyond 10,000 Neurons with Standard Two-Photon Microscopy

    bioRxiv.

    https://doi.org/10.1101/061507

    • Google Scholar
24. 1. Palmer LM
    2. Shai AS
    3. Reeve JE
    4. Anderson HL
    5. Paulsen O
    6. Larkum ME

    (2014) NMDA spikes enhance action potential generation during sensory input

    Nature Neuroscience 17:383–390.

    https://doi.org/10.1038/nn.3646

    • PubMed
    • Google Scholar
25. 1. Papadopoulos PM
    2. Katz MJ
    3. Bruno G

    (2003) NPACI rocks: tools and techniques for easily deploying manageable linux clusters

    Concurrency and Computation 15:707–725.

    https://doi.org/10.1002/cpe.722

    • Google Scholar
26. 1. Park J
    2. Papoutsi A
    3. Ash RT
    4. Marin MA
    5. Poirazi P
    6. Smirnakis SM

    (2019) Contribution of apical and basal dendrites to orientation encoding in mouse V1 L2/3 pyramidal neurons

    Nature Communications 10:5372.

    https://doi.org/10.1038/s41467-019-13029-0

    • PubMed
    • Google Scholar
27. Software

    1. Petousakis EK

    (2023) Petousakis_Etal_2023, version swh:1:rev:c6d1b45694ecb5cd16c8e900d0012ab9f647b598

    Software Heritage.

    https://archive.softwareheritage.org/swh:1:dir:c861eb307043be8183ee29686bcf2d89a2711a05;origin=https://github.com/kepetousakis/petousakis_etal_2023;visit=swh:1:snp:6d9354f290ef5d91c737aef1104917f2820c35f2;anchor=swh:1:rev:c6d1b45694ecb5cd16c8e900d0012ab9f647b598
28. 1. Poleg-Polsky A

    (2019) Dendritic spikes expand the range of well tolerated population noise structures

    The Journal of Neuroscience 39:9173–9184.

    https://doi.org/10.1523/JNEUROSCI.0638-19.2019

    • PubMed
    • Google Scholar
29. 1. Rao RPN
    2. Ballard DH

    (1999) Predictive coding in the visual cortex: a functional interpretation of some extra-classical receptive-field effects

    Nature Neuroscience 2:79–87.

    https://doi.org/10.1038/4580

    • PubMed
    • Google Scholar
30. 1. Rossi LF
    2. Harris KD
    3. Carandini M

    (2020) Spatial connectivity matches direction selectivity in visual cortex

    Nature 588:648–652.

    https://doi.org/10.1038/s41586-020-2894-4

    • PubMed
    • Google Scholar
31. 1. Roth MM
    2. Dahmen JC
    3. Muir DR
    4. Imhof F
    5. Martini FJ
    6. Hofer SB

    (2016) Thalamic nuclei convey diverse contextual information to layer 1 of visual cortex

    Nature Neuroscience 19:299–307.

    https://doi.org/10.1038/nn.4197

    • PubMed
    • Google Scholar
32. 1. Schindelin J
    2. Arganda-Carreras I
    3. Frise E
    4. Kaynig V
    5. Longair M
    6. Pietzsch T
    7. Preibisch S
    8. Rueden C
    9. Saalfeld S
    10. Schmid B
    11. Tinevez JY
    12. White DJ
    13. Hartenstein V
    14. Eliceiri K
    15. Tomancak P
    16. Cardona A

    (2012) Fiji: an open-source platform for biological-image analysis

    Nature Methods 9:676–682.

    https://doi.org/10.1038/nmeth.2019

    • PubMed
    • Google Scholar
33. 1. Scholl B
    2. Tan AYY
    3. Corey J
    4. Priebe NJ

    (2013) Emergence of orientation selectivity in the mammalian visual pathway

    The Journal of Neuroscience 33:10616–10624.

    https://doi.org/10.1523/JNEUROSCI.0404-13.2013

    • PubMed
    • Google Scholar
34. 1. Schuz A
    2. Palm G

    (1989)

    Density of neurons and synapses

    Journal of Comparative Neurology 455:442–455.

    • Google Scholar
35. 1. Simons DJ
    2. Chabris CF

    (1999) Gorillas in our midst: sustained inattentional blindness for dynamic events

    Perception 28:1059–1074.

    https://doi.org/10.1068/p281059

    • PubMed
    • Google Scholar
36. 1. Smith SL
    2. Smith IT
    3. Branco T
    4. Häusser M

    (2013) Dendritic spikes enhance stimulus selectivity in cortical neurons in vivo

    Nature 503:115–120.

    https://doi.org/10.1038/nature12600

    • PubMed
    • Google Scholar
37. 1. Spruston N

    (2008) Pyramidal neurons: dendritic structure and synaptic integration

    Nature Reviews Neuroscience 9:206–221.

    https://doi.org/10.1038/nrn2286

    • PubMed
    • Google Scholar
38. 1. van Elburg RAJ
    2. van Ooyen A

    (2010) Impact of dendritic size and dendritic topology on burst firing in pyramidal cells

    PLOS Computational Biology 6:e1000781.

    https://doi.org/10.1371/journal.pcbi.1000781

    • PubMed
    • Google Scholar
39. 1. Wilson DE
    2. Whitney DE
    3. Scholl B
    4. Fitzpatrick D

    (2016) Orientation selectivity and the functional clustering of synaptic inputs in primary visual cortex

    Nature Neuroscience 19:1003–1009.

    https://doi.org/10.1038/nn.4323

    • PubMed
    • Google Scholar

Author details

1. Konstantinos-Evangelos Petousakis

   1. Department of Biology, University of Crete, Heraklion, Greece
   2. IMBB, FORTH, Heraklion, Greece

   Contribution

   Data curation, Software, Formal analysis, Validation, Visualization, Methodology, Writing – original draft

   Contributed equally with

   Jiyoung Park

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-2022-1671

2. Jiyoung Park

   Department of Neurology, Brigham and Women’s Hospital and Jamaica Plain Veterans Administration Hospital, Harvard Medical School, Boston, United States

   Contribution

   Resources, Data curation, Writing – review and editing

   Contributed equally with

   Konstantinos-Evangelos Petousakis

   Competing interests

   No competing interests declared

3. Athanasia Papoutsi

   IMBB, FORTH, Heraklion, Greece

   Contribution

   Conceptualization, Supervision, Writing – original draft, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-2466-7259

4. Stelios Smirnakis

   Department of Neurology, Brigham and Women’s Hospital and Jamaica Plain Veterans Administration Hospital, Harvard Medical School, Boston, United States

   Contribution

   Conceptualization, Resources, Supervision, Writing – review and editing

   For correspondence

   smsmirnakis@bwh.harvard.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-1929-2811

5. Panayiota Poirazi

   IMBB, FORTH, Heraklion, Greece

   Contribution

   Conceptualization, Supervision, Funding acquisition, Writing – original draft, Project administration, Writing – review and editing

   For correspondence

   poirazi@imbb.forth.gr

   Competing interests

   Reviewing editor, eLife

   "This ORCID iD identifies the author of this article:" 0000-0001-6152-595X

• 942

  views

• 152

  downloads

• 6

  citations

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