CoChat
Sign Up
PreprintCited by 25Neuroscience and Neural EngineeringTactile and Sensory InteractionsEEG and Brain-Computer Interfaces

Optimization of Neuroprosthetic Vision via End-to-end Deep Reinforcement Learning

Open access
Küçükoğlu, Burcu, Rueckauer, Bodo, Ahmad, Nasir, de Ruyter van Steveninck, Jaap, Güçlü, Umut, van Gerven, Marcel·2022·International Journal of Neural Systems
DOI 10.1101/2022.02.25.482017

Preprint — published as 10.1142/s0129065722500526

Open this paper in CoChat — free account

Abstract

Visual neuroprostheses are a promising approach to restore basic sight in visually impaired people. A major challenge is to condense the sensory information contained in a complex environment into meaningful stimulation patterns at low spatial and temporal resolution. Previous approaches considered task-agnostic feature extractors such as edge detectors or semantic segmentation, which are likely suboptimal for specific tasks in complex dynamic environments. As an alternative approach, we propose to optimize stimulation patterns by end-to-end training of a feature extractor using deep reinforcement learning agents in virtual environments. We present a task-oriented evaluation framework to compare different stimulus generation mechanisms, such as static edge-based and adaptive end-to-end approaches like the one introduced here. Our experiments in Atari games show that stimulation patterns obtained via task-dependent end-to-end optimized reinforcement learning result in equivalent or improved performance compared to fixed feature extractors on high difficulty levels. These findings signify the relevance of adaptive reinforcement learning for neuroprosthetic vision in complex environments.

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

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

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

Integrity signals

Registry

Crossref · 10.1101/2022.02.25.482017

Cites

34 of 34 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

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

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

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?

  • This kind of AI learning is often performed in virtual environments due to the lack of real data, but current methods have employed virtual platforms with low realism, limiting their real-world usability.

    “This kind of AI learning is often performed in virtual environments due to the lack of real data, but current methods have employed virtual platforms with low realism, limiting their real-world usability.87,89,90”

    From Visual Prostheses in the Era of Artificial Intelligence Technology (2025) · cites this paper as [89]

    The Experiments and Results section (p. 7) says the authors trained their model on self-generated frames from simulated Atari environments. The Conclusion (p. 14) acknowledges that Atari games may seem distant from typical real-world situations and leaves evaluation in naturalistic environments for future work. The paper does not say that a lack of real data drove the use of virtual environments or that low realism limits the real-world usability of current methods. (LLM, conf=0.84)

    Experimental — accuracy not yet measured

  • Of the five end-to-end studies identified, two relied on auto-encoding strategies and three entailed reinforcement learning approaches.

    “Of the five end-to-end studies that we identified, two relied on auto-encoding strategies,88,91 where the task to be learnt is reconstructing an original image, and three entailed reinforcement learning approaches,87,89,90 for which the scope of learned tasks is wider.”

    From Visual Prostheses in the Era of Artificial Intelligence Technology (2025) · cites this paper as [89]

Self-check

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

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

Cites

Flagged first, then by citations

Research works printed but not linked: 5

  1. [14]14. A. Perez-Yus, J. Bermudez-Cameo, G. Lopez-Nicolas and J. J. Guerrero, Depth and motion cues with phosphene patterns for prosthetic vision, in Proc. IEEE Int. Conf. Computer Vision ( ICCV ) Workshops (Venice, Italy, 2017).
  2. [25]25. C. Zhang, O. Vinyals, R. Munos and S. Bengio, arXiv:abs/1804.06893.
  3. [32]32. J. Schulman, F. Wolski, P. Dhariwal, A. Radford and O. Klimov, arXiv:1707.06347. · arXiv
  4. [33]33. G. Bradski, The OpenCV library, Dr. Dobb's Journal of Software Tools 25 (2000) 120-125.
  5. [38]38. A. P. Badia, B. Piot, S. Kapturowski, P. Sprechmann, A. Vitvitskyi, Z. D. Guo and C. Blundell, Agent57: Outperforming the Atari human benchmark, in Proc. 37 th Int. Conf. Machine Learning , Proceedings of Machine Learning Research, Vol. 119 13-18 July 2020, eds. H. D. III and A. Singh, pp. 507-517.

Cited by

Sorted by citations

Related

Authors
Küçükoğlu, B., Rueckauer, B., Ahmad, N., de Ruyter van Steveninck, J., Güçlü, U., & van Gerven, M.
Affiliations
Radboud University Nijmegen (NL)
Corresponding author
Küçükoğlu, Burcu
ORCID
Venue
International Journal of Neural Systems
Publisher
openRxiv
Published
2022-02-28
Volume
32
Issue
11
Pages
2250052-2250052
ISSN
0129-0657 (ISSN-L)
DOI
10.1101/2022.02.25.482017
OpenAlex
W4289171906
Cited by
25
References
33
Citations by year
2026: 2 · 2025: 7 · 2024: 2 · 2023: 12 · 2022: 2
FWCI
2.54
Type
journal article
Language
en
Open access
Licence
cc-by
Locations
  • International Journal of Neural Systems (publishedVersion, cc-by)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion)
  • PubMed (publishedVersion)
  • Radboud Repository (Radboud University) (submittedVersion, other-oa)
Funders
European Commission (899287, HORIZON2020)
Topics
  • Life Sciences › Neuroscience › Cellular and Molecular Neuroscience › Neuroscience and Neural Engineering
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Tactile and Sensory Interactions
  • Life Sciences › Neuroscience › Cognitive Neuroscience › EEG and Brain-Computer Interfaces
Keywords
end-to-end reinforcement learning, feature extraction, visually impaired people, virtual worlds, Atari games, task-oriented evaluation, adaptive stimulation
MeSH terms
Humans, Learning, Reinforcement, Psychology

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