Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 23 of the 23 references it cites were found in a registry or an index.

Of maps and grids

Open access
Grasso, Matteo, Haun, Andrew M, Tononi, Giulio · 2021 · Neuroscience of Consciousness

Abstract

Neuroscience has made remarkable advances in accounting for how the brain performs its various functions. Consciousness, too, is usually approached in functional terms: the goal is to understand how the brain represents information, accesses that information, and acts on it. While useful for prediction, this functional, information-processing approach leaves out the subjective structure of experience: it does not account for how experience feels. Here, we consider a simple model of how a “grid-like” network meant to resemble posterior cortical areas can represent spatial information and act on it to perform a simple “fixation” function. Using standard neuroscience tools, we show how the model represents topographically the retinal position of a stimulus and triggers eye muscles to fixate or follow it. Encoding, decoding, and tuning functions of model units illustrate the working of the model in a way that fully explains what the model does. However, these functional properties have nothing to say about the fact that a human fixating a stimulus would also “see” it—experience it at a location in space. Using the tools of Integrated Information Theory, we then show how the subjective properties of experienced space—its extendedness—can be accounted for in objective, neuroscientific terms by the “cause-effect structure” specified by the grid-like cortical area. By contrast, a “map-like” network without lateral connections, meant to resemble a pretectal circuit, is functionally equivalent to the grid-like system with respect to representation, action, and fixation but cannot account for the phenomenal properties of space.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1093/nc/niab022

Cites

23 of 23 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1. The fixation network. (A) The receptor layer (the rectangles) is enclosed in an eye with a narrow aperture. The receptors send outputs to the brain, which sends outputs to muscles that move the eye. The brain will have other regions that carry out other functions, but, here, we are interested only in the spatial “fixation” function that is carried out by the three layers shown explicitly in (B). (B) Grid and map versions of the network. The input layer of each network is the same: an array of receptors (a–g) that send their sole output to a unit in the middle layer. The middle layer of both systems (the “brain”) consists of seven self-connected units (A–G), each of which receives a single external input from one receptor. Each of these units sends outputs to several muscle units (h–k) in the top layer. No layer of the network sends outputs to any earlier layer. The grid and map are distinguished by the presence of lateral connections in the grid. (C) Activation functions for each unit type.
Figure 1. The fixation network. (A) The receptor layer (the rectangles) is enclosed in an eye with a narrow aperture. The receptors send outputs to the brain, which sends outputs to muscles that move the eye. The brain will have other regions that carry out other functions, but, here, we are interested only in the spatial “fixation” function that is carried out by the three layers shown explicitly in (B). (B) Grid and map versions of the network. The input layer of each network is the same: an array of receptors (a–g) that send their sole output to a unit in the middle layer. The middle layer of both systems (the “brain”) consists of seven self-connected units (A–G), each of which receives a single external input from one receptor. Each of these units sends outputs to several muscle units (h–k) in the top layer. No layer of the network sends outputs to any earlier layer. The grid and map are distinguished by the presence of lateral connections in the grid. (C) Activation functions for each unit type.CC BY · PubMed Central
Figure 2. Foveation function. The grid/map system is foveating a dark spot on a dark screen. To the right of the foveated spot, a bright dot stimulates one off-center receptor (“b”) in the eye. This produces an imbalance of motor activity that will rotate the eye to the right so that it will come to foveate the bright dot. Active connections are drawn in red, and inactive connections in black.
Figure 2. Foveation function. The grid/map system is foveating a dark spot on a dark screen. To the right of the foveated spot, a bright dot stimulates one off-center receptor (“b”) in the eye. This produces an imbalance of motor activity that will rotate the eye to the right so that it will come to foveate the bright dot. Active connections are drawn in red, and inactive connections in black.CC BY · PubMed Central
Figure 3. Fixation in the grid and map networks. A target stimulus was placed at the outer edge of the receptor array, just in view of receptor “a.” Each receptor is a bin, the space between the horizontal grid lines. Each gray line traces an individual simulation of the eye’s response to the target: the stimulus position never changes, but the eye moves to fixate it, meaning that the stimulus comes to fall on receptor “d.” The red lines indicate the median of each system over 2000 simulations. The grid and the map behave very similarly—on the left, the grid and map median performance is shown to be indistinguishable.
Figure 3. Fixation in the grid and map networks. A target stimulus was placed at the outer edge of the receptor array, just in view of receptor “a.” Each receptor is a bin, the space between the horizontal grid lines. Each gray line traces an individual simulation of the eye’s response to the target: the stimulus position never changes, but the eye moves to fixate it, meaning that the stimulus comes to fall on receptor “d.” The red lines indicate the median of each system over 2000 simulations. The grid and the map behave very similarly—on the left, the grid and map median performance is shown to be indistinguishable.CC BY · PubMed Central
Figure 4. Retinotopy, encoding, and decoding in the grid and map systems. Both the grid and map networks are retinotopic maps of the input layer. (A) The retinotopy stimulus was a bright dot shown to each receptor for 16 time steps, followed by 16 steps of blank display, followed by a stimulus at the next position, etc. When the stimulus had crossed from one end of the retina to the other, it repeated in the opposite direction. The whole 512 time step stimulus was repeated 16 times for each of 20 identical grid or map networks, roughly emulating the parameters of an fMRI retinotopy experiment (with its TRs, trials and repeats, and many subjects). Throughout this experiment, the eye’s muscles were disabled. (B) Grid and (C) map respond similarly to the stimulus. Colors reflect the average proportion of activations over the 320 trials. (D) and (E) show how the stimulus is encoded into grid or map: if an input unit is ON, its target in the grid or map has about P = 0.5 of being ON in the next time step. (F) and (G) show how the input can be decoded from grid or map responses: if a grid or map unit is ON, then there is about P = 0.85 that its input unit was ON in the previous time step.
Figure 4. Retinotopy, encoding, and decoding in the grid and map systems. Both the grid and map networks are retinotopic maps of the input layer. (A) The retinotopy stimulus was a bright dot shown to each receptor for 16 time steps, followed by 16 steps of blank display, followed by a stimulus at the next position, etc. When the stimulus had crossed from one end of the retina to the other, it repeated in the opposite direction. The whole 512 time step stimulus was repeated 16 times for each of 20 identical grid or map networks, roughly emulating the parameters of an fMRI retinotopy experiment (with its TRs, trials and repeats, and many subjects). Throughout this experiment, the eye’s muscles were disabled. (B) Grid and (C) map respond similarly to the stimulus. Colors reflect the average proportion of activations over the 320 trials. (D) and (E) show how the stimulus is encoded into grid or map: if an input unit is ON, its target in the grid or map has about P = 0.5 of being ON in the next time step. (F) and (G) show how the input can be decoded from grid or map responses: if a grid or map unit is ON, then there is about P = 0.85 that its input unit was ON in the previous time step.CC BY · PubMed Central
Figure 5. Input/output (i/o) functional descriptions. Each system was run through 10 000 iterations of the “drifting target” condition for 200 time steps each. This i/o description emulates a “perturb each input unit and observe average system response” experiment. Each matrix (A, B) shows the probability of some output state (of the muscle units, listed on the y-axis) given some single-unit-ON input state (of the receptor units, on the x-axis). Not all output states are possible for either system (e.g. HIJK = [0101] is not possible); only possible states are listed. (C) “Tuning functions” on the input array, given each possible output state. Five of the seven possible output states are associated with a selective tuning function on the input. Grid (black lines) and map (red dashed lines) systems hardly differ in this property.
Figure 5. Input/output (i/o) functional descriptions. Each system was run through 10 000 iterations of the “drifting target” condition for 200 time steps each. This i/o description emulates a “perturb each input unit and observe average system response” experiment. Each matrix (A, B) shows the probability of some output state (of the muscle units, listed on the y-axis) given some single-unit-ON input state (of the receptor units, on the x-axis). Not all output states are possible for either system (e.g. HIJK = [0101] is not possible); only possible states are listed. (C) “Tuning functions” on the input array, given each possible output state. Five of the seven possible output states are associated with a selective tuning function on the input. Grid (black lines) and map (red dashed lines) systems hardly differ in this property.CC BY · PubMed Central
Figure 6. Phenomenal space. The phenomenal experience of visual space is of a canvas filled with spots (distinctions) of any size, related among themselves in a specific “spatial” way: every spot overlaps itself and overlaps with other spots according to connection, fusion, and inclusion (see text and Haun and Tononi (2019) for further explanation).
Figure 6. Phenomenal space. The phenomenal experience of visual space is of a canvas filled with spots (distinctions) of any size, related among themselves in a specific “spatial” way: every spot overlaps itself and overlaps with other spots according to connection, fusion, and inclusion (see text and Haun and Tononi (2019) for further explanation).CC BY · PubMed Central
Figure 7. Cause-effect structures of grid and map. (A) The grid specifies a single cause-effect structure (indicated by the single dashed line) composed of 79 causal distinctions (“spots”) and a large number of relations among them. Here, only a representative subset of the distinctions and relations composing the structure is shown. Cyan edges indicate that two distinctions are related by connection and fusion; purple edges indicate that one distinction includes the other. The structure suggested here is that specified by the “all OFF” grid. The state of the system does affect the existence of the grid’s distinctions (Haun and Tononi 2019). (B) The map specifies seven independent, minimal cause-effect structures (indicated by the seven dashed outlines) each composed of a single first-order distinction.
Figure 7. Cause-effect structures of grid and map. (A) The grid specifies a single cause-effect structure (indicated by the single dashed line) composed of 79 causal distinctions (“spots”) and a large number of relations among them. Here, only a representative subset of the distinctions and relations composing the structure is shown. Cyan edges indicate that two distinctions are related by connection and fusion; purple edges indicate that one distinction includes the other. The structure suggested here is that specified by the “all OFF” grid. The state of the system does affect the existence of the grid’s distinctions (Haun and Tononi 2019). (B) The map specifies seven independent, minimal cause-effect structures (indicated by the seven dashed outlines) each composed of a single first-order distinction.CC BY · PubMed Central
Figure 8. Extendedness. Phenomenal spots are represented in the top left of every panel; the corresponding distinctions are represented in the rest of the panel. For simplicity, spots are represented as 2D, although the space specified by a 1D grid would be 1D. (A) Phenomenal spatial distinctions (“spots”) are reflexive: they overlap or “point to” themselves. A causal distinction from the grid (here, the system is in the all OFF state) also overlaps itself, yielding a reflexive relation. (B) Spots connect to one another: when they overlap partially, their intersection is also a spot. When causal distinctions from a grid overlap partially, another distinction exists corresponding to the intersection. (C) Spots fuse with one another: when they overlap, their union is also a spot. When causal distinctions from a grid overlap, another distinction exists corresponding to their fusion. (D) Spots include one another in the sense that one spot might fully overlap another as well as some distinct spot. Distinctions from a grid often overlap in a corresponding way. The overlapping spots inset in each panel are labeled with the analogous distinction in the example, but it should be clear that the space of the line grid is 1D, not 2D like the phenomenal space of Fig. 6.
Figure 8. Extendedness. Phenomenal spots are represented in the top left of every panel; the corresponding distinctions are represented in the rest of the panel. For simplicity, spots are represented as 2D, although the space specified by a 1D grid would be 1D. (A) Phenomenal spatial distinctions (“spots”) are reflexive: they overlap or “point to” themselves. A causal distinction from the grid (here, the system is in the all OFF state) also overlaps itself, yielding a reflexive relation. (B) Spots connect to one another: when they overlap partially, their intersection is also a spot. When causal distinctions from a grid overlap partially, another distinction exists corresponding to the intersection. (C) Spots fuse with one another: when they overlap, their union is also a spot. When causal distinctions from a grid overlap, another distinction exists corresponding to their fusion. (D) Spots include one another in the sense that one spot might fully overlap another as well as some distinct spot. Distinctions from a grid often overlap in a corresponding way. The overlapping spots inset in each panel are labeled with the analogous distinction in the example, but it should be clear that the space of the line grid is 1D, not 2D like the phenomenal space of Fig. 6.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 23 references

Cited by

See all 25

Metadata

Authors
Grasso, M., Haun, A. M., & Tononi, G.
Affiliations
University of Wisconsin–Madison (US)
Corresponding author
Tononi, Giulio
Venue
Neuroscience of Consciousness
Published
2021-09-01
Volume
2021
Issue
2
Pages
niab022-niab022
ISSN
2057-2107 (ISSN-L)
OpenAlex
W3201266294
Cited by
19
References
23
Citations by year
2026: 5 · 2025: 4 · 2024: 3 · 2023: 2 · 2022: 3 · 2021: 2
FWCI
2.42
Type
journal article
Language
en
Open access
Locations
  • Neuroscience of Consciousness (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, other-oa)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
Funders
  • Tiny Blue Dot Foundation
  • Templeton World Charity Foundation
Topics
  • Social Sciences › Arts and Humanities › Philosophy › Philosophy and Theoretical Science
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Neural dynamics and brain function
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Visual perception and processing mechanisms
Keywords
grid cells, eye fixations, spatial representation, integrated information theory, phenomenal space, subjective experience, causal structure, topographic map

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Of maps and grids. https://app.cochat.ai/papers/10.1093/nc/niab022
The paper
Grasso, M., Haun, A. M., & Tononi, G. (2021). Of maps and grids. Neuroscience of Consciousness. https://doi.org/10.1093/nc/niab022

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.