Number detectors spontaneously emerge in a deep neural network designed for visual object recognition
Open accessThe DOI matches its Crossref record; 45 of the 45 references it cites were found in a registry or an index.
Abstract
A deep neural network trained only on visual object recognition develops tuned number detectors reminiscent of real neurons.
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.
Integrity signals
Registry
Cites
Statistics
Figures

![Fig. 2 Numerosity-tuned units emerging in the HCNN. (A) Examples of the stimuli used to assess numerosity encoding. Standard stimuli contain dots of the same average radius. Dots in Area & Density stimuli have a constant total area and density across all numerosities. Dots in Shape & Convex hull stimuli have random shapes and a uniform pentagon convex hull (for numerosities >4). (B) Tuning curves for individual numerosity-selective network units. Colored curves show the average responses for each stimulus set. Black curves show the average responses over all stimulus set. Error bars indicate SE measure. PN, preferred numerosity. (C) Same as (B), but for neurons in monkey prefrontal cortex (20). Only the average responses over all stimulus sets are shown. (D) Distribution of preferred numerosities of the numerosity-selective network units. (E) Same as (D), but for real neurons recorded in monkey prefrontal cortex [data from (20)].](https://figures.cochat.ai/paper-figures/aaf715d51e0d474b8c325f99b3eb8af1/number-detectors-spontaneously-emerge-in-a-deep-neural-network-designed-for-figure-2.jpg)

![Fig. 4 Tuning properties of numerosity-selective network units. (A) Left: Average tuning curves for network units preferring each numerosity plotted on a linear scale. Right: Same tuning curves plotted on a logarithmic scale. (B) Average goodness-of-fit measure for fitting Gaussian functions to the tuning curves on different scales [Plinear-log = 0.009; Plinear-pow(1/2) = 0.003; Plinear-pow(1/3) = 0.001]. (C) SD of the best-fitting Gaussian function for each of the tuning curves of numerosity-selective network units for different scales.](https://figures.cochat.ai/paper-figures/627b66a232dc435c8d7b371f03af87de/number-detectors-spontaneously-emerge-in-a-deep-neural-network-designed-for-figure-4.jpg)
![Fig. 5 Relevance of numerosity-selective units to network performance. Average activity of numerosity-selective network units shown as a function of numerical distance between preferred numerosities and sample numerosities in the matching task. Left: Data from network units. Responses were average separately for correct trials (black) and error trials (gray). Responses during error trials are normalized to the maximum average response during correct trials (P = 0.019). Right: Same plot but for real neurons recorded from monkey prefrontal cortex [data from (20)].](https://figures.cochat.ai/paper-figures/ee8c0e1e56fb4bdbaaa7270a62d0c085/number-detectors-spontaneously-emerge-in-a-deep-neural-network-designed-for-figure-5.jpg)
![Fig. 6 Performance of the HCNN model in the numerosity matching task. (A) Left: Performance functions resulting from the discrimination of numerosities plotted on a linear scale. Each curve shows the probability of the model predicting that the sample image contains the same number of items as the test image (peak of the function). Sample numerosity is indicated above each curve. Right: Same performance functions but plotted on a logarithmic scale. (B) Average goodness-of-fit measure for fitting Gaussian functions to the performance tuning curves on different scales [Plinear-log = 0.003; Plinear-pow(1/2) = 0.049; Plinear-pow(1/3) = 0.016]. *P < 0.05. (C) SD of the best-fitting Gaussian function for each of the performance tuning curves for different scales.](https://figures.cochat.ai/paper-figures/f8bf766d1b534ddcaa16bedda91e1a87/number-detectors-spontaneously-emerge-in-a-deep-neural-network-designed-for-figure-6.jpg)
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.
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?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Models trained to process images, for example, spontaneously learned how to count (30, 31) and differentially process central and peripheral image areas (32), as well as experience human-like optical illusions (33).”
From Evaluating large language models in theory of mind tasks (2024) · cites this paper as [30]
Experimental — accuracy not yet measured
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
Very Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 114,148 · cited 1×
“Here, we use a hierarchical convolutional neural network (HCNN), a class of biologically inspired models that have recently achieved great success in computer vision applications (14, 15) and in the modeling of the ventral visual stream (16, 17).”
MethodImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155 · cited 4×
“Network units that had the same receptive fields in convolutional layers competed with each other using a simple form of lateral inhibition (14).”
BackgroundMethodInfluentialImageNet Large Scale Visual Recognition Challenge
Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., Huang, Z., Karpathy, A., Khosla, A., Bernstein, M., et al. (2015) · International Journal of Computer Vision · doi:10.1007/s11263-015-0816-y · Cited by 39,380
(2014) · Cited by 33,419
Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Ioffe, S., & Szegedy, C. (2015) · arXiv · doi:10.48550/arxiv.1502.03167 · Cited by 27,818 · cited 1×
“Before applying the activation function, the outputs of the convolution operation were normalized to have zero mean and unit SD [batch normalization; (40)].”
Method(2010) · Cited by 14,134
Receptive fields, binocular interaction and functional architecture in the cat’s visual cortex
Hubel, D. H., & Wiesel, T. N. (1962) · The Journal of Physiology · doi:10.1113/jphysiol.1962.sp006837 · Cited by 13,680 · cited 1×
“The network model was an instance of HCNNs (18), originally inspired by the discovery of simple and complex cells in early visual cortex (19).”
Method(2017) · Cited by 11,625
An overview of gradient descent optimization algorithms
Ruder, S. (2016) · arXiv · doi:10.48550/arxiv.1609.04747 · Cited by 4,962 · cited 1×
“The minimization was performed using mini-batch gradient descent (43) with a batch size of 256 images, a learning rate of 0.”
MethodPerformance-optimized hierarchical models predict neural responses in higher visual cortex
Yamins, D. L. K., Hong, H., Cadieu, C. F., Solomon, E. A., Seibert, D., & DiCarlo, J. J. (2014) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1403112111 · Cited by 1,924
Tuning Curves for Approximate Numerosity in the Human Intraparietal Sulcus
Piazza, M., Izard, V., Pinel, P., Le Bihan, D., & Dehaene, S. (2004) · Neuron · doi:10.1016/j.neuron.2004.10.014 · Cited by 1,099 · cited 2×
“Such a simultaneous assessment of the number of items is themost common presentation format to investigate the nonsymbolic number sense in humans (5, 6, 24, 25, 29) and animals (24, 34, 35), and our deepnetwork spontaneously derived the number of items from these multi-dot patterns.”
BackgroundMethodRepresentation of Number in the Brain
Nieder, A., & Dehaene, S. (2009) · Annual Review of Neuroscience · doi:10.1146/annurev.neuro.051508.135550 · Cited by 847 · cited 1×
“Because adjacent numerical values in dot displays are difficult to discriminate for humans and monkeys due to the numerical distance effect (12, 32), the network is not expected to reach perfect discrimination.”
BackgroundSelf-citationCadieu, C. F., Hong, H., Yamins, D. L. K., Pinto, N., Ardila, D., Solomon, E. A., Majaj, N. J., & DiCarlo, J. J. (2014) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1003963 · Cited by 824
Newborn infants perceive abstract numbers
Izard, V., Sann, C., Spelke, E. S., & Streri, A. (2009) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.0812142106 · Cited by 778 · cited 2×
“This capacity allows newborn human infants (3) and animals (4) to assess the number of items in a visual scene.”
BackgroundLe Corre, M., & Carey, S. (2007) · Cognition · doi:10.1016/j.cognition.2006.10.005 · Cited by 731
Representation of the Quantity of Visual Items in the Primate Prefrontal Cortex
Nieder, A., Freedman, D. J., & Miller, E. K. (2002) · Science · doi:10.1126/science.1072493 · Cited by 699 · cited 2×
“Analogous to the approach used to detect numerosityselective neurons in monkeys and humans (11, 20, 31), a two-way ANOVA with numerosity (16 numerosities) and stimulus sets (three numerosity sets: one standard and two control) as factors was applied to find network units that exhibited a significant effect for numeros…”
MethodResultSelf-citationOrdering of the Numerosities 1 to 9 by Monkeys
Brannon, E. M., & Terrace, H. S. (1998) · Science · doi:10.1126/science.282.5389.746 · Cited by 696
Development of Elementary Numerical Abilities: A Neuronal Model
Dehaene, S., & Changeux, J.-P. (1993) · Journal of Cognitive Neuroscience · doi:10.1162/jocn.1993.5.4.390 · Cited by 695 · cited 4×
“…trained for object classification, the network was presented with images depicting different numbers of items (numerosities 1, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, and 30), and the units ’ responses in the final layer of the feature extraction network (see Table 1) were analyzed.”
MethodInfluentialBurr, D., & Ross, J. (2008) · Current Biology · doi:10.1016/j.cub.2008.02.052 · Cited by 592
Nieder, A., & Miller, E. K. (2003) · Neuron · doi:10.1016/s0896-6273(02)01144-3 · Cited by 512 · cited 6×
“If a scale is suited to the tuning curves, they should become symmetric around preferred numerosities when plotted on that scale, and therefore, the goodness of fit (r(2) score) of the Gaussian function to the tuning curves should be increased (21).”
BackgroundMethodResultInfluentialSelf-citationRepresentation of Number in Animals and Humans: A Neural Model
Verguts, T., & Fias, W. (2004) · Journal of Cognitive Neuroscience · doi:10.1162/0898929042568497 · Cited by 454 · cited 3×
“Unlike previous models for visual number coding that relied on either hard-coded connection strengths (22, 23) or training on non-naturalistic dot pattern stimuli (13), our model was merely trained to classify natural images in a task that was unrelated to numerosity.”
BackgroundMethodInfluentialNieder, A. (2016) · Nature Reviews Neuroscience · doi:10.1038/nrn.2016.40 · Cited by 385 · cited 5×
“Because adjacent numerical values in dot displays are difficult to discriminate for humans and monkeys due to the numerical distance effect (12, 32), the network is not expected to reach perfect discrimination.”
BackgroundMethodResultInfluentialSelf-citationComparisons of digits and dot patterns.
Buckley, P. B., & Gillman, C. B. (1974) · Journal of Experimental Psychology · doi:10.1037/h0037361 · Cited by 376
Number; the Language of Science.
Dantzig, T. (1930) · The Economic Journal · doi:10.2307/2224269 · Cited by 354
Emergence of a 'visual number sense' in hierarchical generative models
Stoianov, I., & Zorzi, M. (2012) · Nature Neuroscience · doi:10.1038/nn.2996 · Cited by 300 · cited 4×
“Generative neural networks, a class of deep networks that learn to form an internal model of the sensory input, have been shown to become sensitive to numerosity but could not explain the emergence of real number neurons ( 13 ).”
InfluentialTemporal and Spatial Enumeration Processes in the Primate Parietal Cortex
Nieder, A., Diester, I., & Tudusciuc, O. (2006) · Science · doi:10.1126/science.1130308 · Cited by 293 · cited 1×
“Sequentialenumerationrequiresanassessment ofnumberacrosstime,ratherthanacrossspaceasfordotdisplays,and theneuronalmechanismsbetweenthesetwoprocessesdiffer( 36 , 37 ).”
Self-citationNoisy matrix decomposition via convex relaxation: Optimal rates in high dimensions
Agarwal, A., Negahban, S., & Wainwright, M. J. (2012) · The Annals of Statistics · doi:10.1214/12-aos1000 · Cited by 280
Lin, M., Chen, Q., & Yan, S. (2013) · arXiv · doi:10.48550/arxiv.1312.4400 · Cited by 252
A Labeled-Line Code for Small and Large Numerosities in the Monkey Prefrontal Cortex
Nieder, A., & Merten, K. (2007) · The Journal of Neuroscience · doi:10.1523/jneurosci.1056-07.2007 · Cited by 194
Self-citationSupramodal numerosity selectivity of neurons in primate prefrontal and posterior parietal cortices
Nieder, A. (2012) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1204580109 · Cited by 190 · cited 1×
“Sequential enumeration requires an assessment of number across time, rather than across space as for dot displays, and the neuronal mechanisms between these two processes differ (36, 37).”
BackgroundSelf-citationNeurons selective to the number of visual items in the corvid songbird endbrain
Ditz, H. M., & Nieder, A. (2015) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1504245112 · Cited by 179
Self-citationRapid and Direct Encoding of Numerosity in the Visual Stream
Park, J., DeWind, N. K., Woldorff, M. G., & Brannon, E. M. (2015) · Cerebral Cortex · doi:10.1093/cercor/bhv017 · Cited by 170
Neuronal correlates of a visual “sense of number” in primate parietal and prefrontal cortices
Viswanathan, P., & Nieder, A. (2013) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1308141110 · Cited by 168 · cited 3×
“Just like neurons in the brains of numerically naïve animals, about 10% of the network units exhibited numerosity selectivity (9, 10).”
BackgroundResultSelf-citationSingle Neurons in the Human Brain Encode Numbers
Kutter, E. F., Bostroem, J., Elger, C. E., Mormann, F., & Nieder, A. (2018) · Neuron · doi:10.1016/j.neuron.2018.08.036 · Cited by 165
Self-citationDiscrimination of small numerosities in young chicks.
Rugani, R., Regolin, L., & Vallortigara, G. (2008) · Journal of Experimental Psychology: Animal Behavior Processes · doi:10.1037/0097-7403.34.3.388 · Cited by 164 · cited 1×
“This capacity allows newborn human infants (3) and animals (4) to assess the number of items in a visual scene.”
BackgroundNumerosity representations in crows obey the Weber–Fechner law
Ditz, H. M., & Nieder, A. (2016) · Proceedings of the Royal Society B Biological Sciences · doi:10.1098/rspb.2016.0083 · Cited by 145
Self-citationTuning to non‐symbolic proportions in the human frontoparietal cortex
Jacob, S. N., & Nieder, A. (2009) · European Journal of Neuroscience · doi:10.1111/j.1460-9568.2009.06932.x · Cited by 139
Self-citationPsychophysical evidence for the number sense
Burr, D. C., Anobile, G., & Arrighi, R. (2018) · Philosophical Transactions of the Royal Society B Biological Sciences · doi:10.1098/rstb.2017.0045 · Cited by 119
Compressed Scaling of Abstract Numerosity Representations in Adult Humans and Monkeys
Merten, K., & Nieder, A. (2008) · Journal of Cognitive Neuroscience · doi:10.1162/jocn.2008.21032 · Cited by 99 · cited 7×
“For this purpose, we constructed a numerosity matching task (see Materials and Methods for details) that was comparable to the tasks developed for monkeys and humans (24).”
BackgroundMethodResultInfluentialSelf-citationNeurons in the Endbrain of Numerically Naive Crows Spontaneously Encode Visual Numerosity
Wagener, L., Loconsole, M., Ditz, H. M., & Nieder, A. (2018) · Current Biology · doi:10.1016/j.cub.2018.02.023 · Cited by 78 · cited 3×
“Just like neurons in the brains of numerically naïve animals, about 10% of the network units exhibited numerosity selectivity (9, 10).”
BackgroundResultSelf-citationViswanathan, P., & Nieder, A. (2015) · Current Biology · doi:10.1016/j.cub.2015.03.025 · Cited by 74 · cited 1×
“Only with ample training, the performance of monkeys slightly increased over time (24), and also in agreement with this behavioral enhancement, more andmore selective prefrontal cortex neurons represented numerosity (33).”
Self-citationSensory and Working Memory Representations of Small and Large Numerosities in the Crow Endbrain
Ditz, H. M., & Nieder, A. (2016) · Journal of Neuroscience · doi:10.1523/jneurosci.1521-16.2016 · Cited by 73
Self-citationEffects of adaptation on numerosity decoding in the human brain
Castaldi, E., Aagten‐Murphy, D., Tosetti, M., Burr, D. C., & Morrone, M. C. (2016) · NeuroImage · doi:10.1016/j.neuroimage.2016.09.020 · Cited by 70 · cited 1×
“Human psychophysics (5, 6), brain imaging studies in humans (7, 8), and single-neuron recordings in animals support the direct and automatic assessment of numerosity in the brain.”
BackgroundSexual Dimorphism: Can You Smell the Difference?
Rideout, E. J., & Goodwin, S. F. (2008) · Current Biology · doi:10.1016/j.cub.2008.03.035 · Cited by 55
Numerical assessment in the wild: insights from social carnivores
Benson‐Amram, S., Gilfillan, G., & McComb, K. (2018) · Philosophical Transactions of the Royal Society B Biological Sciences · doi:10.1098/rstb.2016.0508 · Cited by 51 · cited 1×
“This could explain why numerically naïve subjects, such as newborns (3) and animals in the wild (26), are innately endowed with numerical capabilities.”
Background
Research works printed but not linked: 3
- S. Dehaene The Number Sense (Oxford Univ. Press 1999).
- Y. LeCun Y. Bengio Convolutional networks for images speech and time-series in The Handbook of Brain Theory and Neural Networks (MIT Press 1995) vol. 3361 pp. 255–258.
- S. Carey The Origin of Concepts: Oxford Series in Cognitive Development (Oxford Univ. Press 2009).
Cited by
Evaluating large language models in theory of mind tasks
Kosinski, M. (2024) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2405460121 · Cited by 339
“Models trained to process images, for example, spontaneously learned how to count ( 30 , 31 ) and differentially process central and peripheral image areas ( 32 ), as well as experience human - like optical illusions ( 33 ).”
BackgroundUsing artificial neural networks to ask ‘why’ questions of minds and brains
Kanwisher, N., Khosla, M., & Dobs, K. (2023) · Trends in Neurosciences · doi:10.1016/j.tins.2022.12.008 · Cited by 182
“Analyses in this vein have revealed the spontaneous emergence of face-selective [88,89] and number-selective units [90,91] in ANNs.”
BackgroundThe Adaptive Value of Numerical Competence
Nieder, A. (2020) · Trends in Ecology & Evolution · doi:10.1016/j.tree.2020.02.009 · Cited by 162
“In support of a putatively innate ANS, artificial number neurons obeying Weber’s and Fechner’s law spontaneously emerged also in a deep network that was merely trained on visual object recognition, never on numbers [133].”
BackgroundSelf-citationBrain-like functional specialization emerges spontaneously in deep neural networks
Dobs, K., Martinez, J., Kell, A. J. E., & Kanwisher, N. (2022) · Science Advances · doi:10.1126/sciadv.abl8913 · Cited by 129
“…then the same should be true for any computational system, including CNNs. Prior work suggests that this may not necessarily be the case: CNNs optimized for generic object recognition transfer well to many other tasks ( 12 – 15 ), including fine-grained discrimination within a category ( 13 ).”
BackgroundEvaluating Large Language Models in Theory of Mind Tasks
Kosinski, M. (2023) · arXiv · doi:10.48550/arxiv.2302.02083 · Cited by 339
“Models trained to process images, for example, spontaneously learned how to count ( 30 , 31 ) and differentially process central and peripheral image areas ( 32 ), as well as experience human - like optical illusions ( 33 ).”
Background
Metadata
- Authors
- Nasr, K., Viswanathan, P., & Nieder, A.
- Affiliations
- University of Tübingen (DE)
- Corresponding author
- Nieder, Andreas
- Venue
- Science Advances
- Published
- 2019-05-03
- Volume
- 5
- Issue
- 5
- Pages
- eaav7903-eaav7903
- ISSN
- 2375-2548 (ISSN-L)
- PMC
- PMC6506249
- OpenAlex
- W2944629072
- Cited by
- 110
- References
- 41
- Citations by year
- 2026: 4 · 2025: 11 · 2024: 18 · 2023: 19 · 2022: 19 · 2021: 20 · 2020: 15 · 2019: 4
- FWCI
- 11.34
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- cc-by-nc
- Locations
- Science Advances (publishedVersion, cc-by-nc)
- PubMed Central (publishedVersion, cc-by-nc)
- PubMed (publishedVersion)
- Funders
- Deutsche Forschungsgemeinschaft (NI 618/10-1)
- Topics
- Social Sciences › Psychology › Developmental and Educational Psychology › Cognitive and developmental aspects of mathematical skills
- Life Sciences › Neuroscience › Cognitive Neuroscience › Visual perception and processing mechanisms
- Social Sciences › Psychology › Experimental and Cognitive Psychology › Multisensory perception and integration
- Keywords
- number sense, numerosity, spontaneous emergence, deep neural network, visual object recognition, number discrimination, Weber-Fechner law, biologically inspired models, visual system
- MeSH terms
- Animals, Humans, Models, Biological, Neurons, Visual Perception, Neural Networks, Computer
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
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.