Neural Networks

About Neural Networks

Neural Networks: Journal Overview

Neural Networks is a monthly peer-reviewed scientific journal published by Elsevier, and the archival journal of three of the world's oldest and most prominent neural-network societies: the International Neural Network Society (INNS), the European Neural Network Society (ENNS), and the Japanese Neural Network Society (JNNS) (sources: sciencedirect.com/journal/neural-networks; inns.org/neural-networks-journal, accessed October 2026). Elsevier's own journal description additionally names the Asia-Pacific Neural Network Society (APNNS) among the affiliated societies — a subscription to the journal is included with membership in each of these societies (source: sciencedirect.com/journal/neural-networks/publish/guide-for-authors).

The journal sits at the intersection of biological and artificial intelligence. It provides a forum for an international community of scholars and practitioners working on all aspects of neural networks — including deep learning and related approaches to artificial intelligence and machine learning. Its remit deliberately spans from cognitive modeling and computational neuroscience, through deep-learning algorithms and mathematical analyses, to engineering and technological applications of systems that significantly use neural-network concepts and learning techniques (source: sciencedirect.com/journal/neural-networks).

This unusually broad range is the journal's defining editorial feature. It is designed to enable the cross-fertilisation of ideas between biological and technological studies, fostering an interdisciplinary community interested in biologically-inspired artificial intelligence. Accordingly, the journal's editorial board brings together experts from psychology, neurobiology, computer science, engineering, mathematics, and physics (source: sciencedirect.com/journal/neural-networks/publish/guide-for-authors). Founded in 1988, Neural Networks has published continuously through the rise of modern deep learning, and today it is one of the longest-running dedicated venues for neural-network research (sources: journal description by Elsevier; inns.org).

History of Neural Networks

Neural Networks was established in 1988, at a moment when connectionist research was undergoing a renaissance driven by backpropagation, Hopfield networks, and renewed interest in parallel distributed processing. The journal's founding editor-in-chief was Stephen Grossberg of Boston University, a pioneering theorist of adaptive resonance theory and one of the founders of the computational neuroscience tradition (sources: the journal's Elsevier description as recorded in standard references; inns.org journal history).

The journal was conceived from the outset as the archival home of the neural-network societies. The International Neural Network Society itself was founded in 1987, and Neural Networks quickly became its flagship publication — a role it has held for nearly four decades. As deep learning emerged in the 2010s, the journal's scope expanded naturally to embrace it: its aims-and-scope text today explicitly names deep learning alongside the field's traditional concerns of cognitive modeling and computational neuroscience (source: sciencedirect.com/journal/neural-networks).

Over its history the journal has been guided by distinguished co-editors-in-chief. An editorial in the journal's pages from 2012 (Doya, Taylor & Wang) documents the stewardship of Kenji Doya, John Taylor, and DeLiang Wang as co-editors-in-chief in that era, during which the journal introduced an expedited review process for its Letters track (source: Neural Networks 25 (2012), via media.journals.elsevier.com). In the current editorial era, the journal is led by Editor-in-Chief Taro Toyoizumi (RIKEN Center for Brain Science) together with co-editors-in-chief from Italy, the United States, and Australia — a leadership spread across three continents that mirrors the journal's international authorship (source: inns.org/neural-networks-journal).

Published twelve times a year by Elsevier, the journal is available electronically through ScienceDirect, and all INNS members receive an online subscription to Neural Networks as part of their membership (source: inns.org/neural-networks-journal). The journal's volume numbering illustrates its steady monthly output: Volume 148 appeared in April 2022 (per INNS's Best Paper Award records), and Volume 202 was in progress on ScienceDirect in October 2026.

Leadership and Editorial Team

According to the International Neural Network Society's official journal page, the current editorial leadership of Neural Networks is (source: inns.org/neural-networks-journal, accessed October 2026):

• Editor-in-Chief: Taro Toyoizumi — RIKEN Center for Brain Science, Wako, Japan. ScienceDirect's journal page likewise names Toyoizumi among the journal's editors.
• Co-Editor-in-Chief: Mauro Forti — University of Siena, Siena, Italy.
• Co-Editor-in-Chief: Robert Kozma — University of Memphis, Tennessee, USA.
• Co-Editor-in-Chief: Tongliang Liu — The University of Sydney, Sydney, Australia.

This team reflects a recent leadership transition at the journal: earlier reference records list DeLiang Wang (The Ohio State University) and Kenji Doya (Okinawa Institute of Science and Technology) as editors-in-chief. Readers should treat the INNS page — the society that owns the journal's archival role — and ScienceDirect as the authoritative current listings (sources: inns.org; sciencedirect.com/journal/neural-networks).

Beyond the chief editors, the journal maintains a broad editorial board whose members span psychology, neurobiology, computer science, engineering, mathematics, and physics — the six disciplines Elsevier's journal description names as the board's constituency (source: sciencedirect.com/journal/neural-networks/publish/guide-for-authors). This interdisciplinary board composition is deliberate: the journal's scope requires reviewers who can judge work ranging from brain modeling to theorem-level mathematical analysis.

Aims & Scope

The journal's official aims-and-scope statement, published on ScienceDirect, presents Neural Networks as a venue for building an international community around neural-network research in its full breadth. Paraphrasing that statement (source: sciencedirect.com/journal/neural-networks/publish/guide-for-authors):

• Community forum: The journal aims to develop and nurture an international community of scholars and practitioners interested in every aspect of neural networks, including deep learning and related AI and machine-learning approaches.
• Full-range coverage: It welcomes work across the entire spectrum of neural-network research — from cognitive modeling and computational neuroscience, through deep-learning algorithms and mathematical analyses, to engineering and technological applications of systems that substantially use neural-network concepts and learning techniques.
• Cross-fertilisation: The breadth is intentional: by publishing biological and technological studies side by side, the journal seeks to cross-fertilise ideas between the two and to grow an interdisciplinary community focused on biologically-inspired artificial intelligence.
• Neural networks must be central: While the scope is wide, submissions are expected to place neural networks at the centre of the work — the journal is not a general AI venue.
• Five sections: Articles are published in one of five sections — Learning Systems; Cognitive Science; Neuroscience; Mathematical and Computational Analysis; and Engineering and Applications.

The INNS society page offers a complementary, slightly differently worded statement of the same scope, emphasising contributions "from behavioral and brain modeling, learning algorithms, through mathematical and computational analyses, to engineering and technological applications" (source: inns.org/neural-networks-journal). In addition to full articles, the journal publishes letters, reviews and tutorials, letters to the editor, editorials, current events.

Editorial Process and Peer Review

Neural Networks is a peer-reviewed journal. Full details of the review workflow, article-type requirements, and author policies are maintained in the journal's Guide for Authors on ScienceDirect, which authors should consult in its current form before submitting. The journal publishes full-length Articles, Letters (short reports for important new results), Reviews and Tutorials, Letters to the Editor, editorials, and current events (article types per the journal's official description).

Speed. Elsevier publishes journal-insight turnaround statistics for Neural Networks on the journal's ScienceDirect page (accessed October 2026): the median time from submission to first decision is 9 days, from submission to decision after review 78 days, from submission to acceptance 208 days, and from acceptance to online publication 8 days. These figures shift over time and should be re-checked on the journal page before relying on them.

Acceptance rate. The journal does not publicly disclose an acceptance rate. Any figure circulating on third-party websites is unofficial and should be treated with caution.

Impact and Metrics

Elsevier's ScienceDirect journal page for Neural Networks does not publish a Journal Impact Factor figure; authors seeking the current impact factor should consult Clarivate's Journal Citation Reports directly. For historical context: the 2022 edition of the Journal Citation Reports (Clarivate, via Web of Science) listed a 2022 Impact Factor of 7.8 for Neural Networks (source: 2022 Journal Citation Reports, Science edition, Clarivate 2023, as cited in the journal's encyclopaedic record). Independent Scopus-based listings place the journal in the Q1 quartile for Artificial Intelligence and Cognitive Neuroscience categories, but these are third-party compilations — verify current rankings in Scopus or SCImago before citing (sources: independent metric aggregators, accessed October 2026).

What is officially documented on the publisher's own pages:

Speed figures from the journal's ScienceDirect page (Elsevier journal insights, accessed October 2026); figures are updated periodically. Indexing per standard bibliographic records.

Submission Guidelines

Manuscripts are submitted electronically through Elsevier's editorial system, linked from the journal's ScienceDirect page. The journal publishes:

• Articles — original, full-length research, assigned to one of the five sections (Learning Systems; Cognitive Science; Neuroscience; Mathematical and Computational Analysis; Engineering and Applications).
• Letters — short reports of important new results warranting rapid publication.
• Reviews or Tutorials — comprehensive topical surveys.
• Letters to the Editor — short communications commenting on previously published papers.
• Editorials and current events.

Detailed formatting, reference-style, figure, ethics, and originality requirements are maintained in the full Guide for Authors on ScienceDirect and change over time — authors should consult the current version before submitting (source: sciencedirect.com/journal/neural-networks/publish/guide-for-authors).

Open Access and APC

Neural Networks is a hybrid journal: authors can publish either through the traditional subscription route or via gold open access. Under the subscription option, no publication fee is charged to authors and published articles are immediately available to subscribers (source: sciencedirect.com/journal/neural-networks).

For the open access route, the journal's ScienceDirect page lists an Article Publishing Charge (APC) of USD 2,950 (excluding taxes), noting that the amount payable may be reduced during submission if the author qualifies for an applicable discount or institutional agreement (source: sciencedirect.com/journal/neural-networks, accessed October 2026). Authors should verify the current APC on the journal's ScienceDirect page at the time of submission, as APCs change over time.

A point of practical value from the society side: INNS members receive a combined 50% discount (25% from the Neural Networks journal and 25% from INNS) on open-access article processing charges, alongside complimentary access to the journal (source: inns.org/neural-networks-journal, accessed October 2026).

Landmark Papers Published in Neural Networks

The journal's standing is reflected in a set of highly influential papers, recognised by the INNS Neural Networks Best Paper Award, an annual award presented at the International Joint Conference on Neural Networks (IJCNN); the next award is to be presented at IJCNN 2026 in Maastricht, the Netherlands, 21–26 June 2026 (source: inns.org/neural-networks-journal). Past awardees listed by INNS include:

• Jürgen Schmidhuber, "Deep Learning in Neural Networks: An Overview" (vol. 61, pp. 85–117, January 2015) — a landmark survey tracing the history and methods of deep learning, and one of the most widely cited papers the journal has ever published; Schmidhuber's own record notes it received the first Best Paper Award ever issued by the journal.
• Steven Grossberg, "Towards solving the hard problem of consciousness: The varieties of brain resonances and the conscious experiences that they support" (vol. 87, pp. 38–95, March 2017) — a major theoretical statement by the journal's founding editor-in-chief on adaptive resonance theory and consciousness.
• German I. Parisi, Ronald Kemker, Jose L. Part, Christopher Kanan & Stefan Wermter, "Continual lifelong learning with neural networks: A review" (vol. 113, pp. 54–71, May 2019) — a field-defining review of continual and lifelong learning.
• Guan'an Wang, Yang Yang, Tianzhu Zhang, Jian Cheng, Zengguang Hou, Prayag Tiwari & Hari Mohan Pandey, "Cross-modality paired-images generation and augmentation for RGB-infrared person re-identification" (vol. 128, pp. 294–304, August 2020).
• Xiao-Lei Zhang, "Multilayer bootstrap networks" (vol. 103, pp. 29–43, July 2018).
• Nikola Kasabov et al., "Evolving Spatio-temporal Data Machines Based on the NeuCube Neuromorphic Framework: Design Methodology and Selected Applications" (vol. 78, pp. 1–14, June 2016).
• Suogui Dang, Yining Wu, Rui Yan & Huajin Tang, "Why grid cells function as a metric for space" (vol. 142, pp. 128–137, October 2021) — connecting neural-network modelling with spatial neuroscience.
• Garrett Bingham & Risto Miikkulainen, "Discovering parametric activation functions" (vol. 148, pp. 48–65, April 2022).

Source for all award citations: the International Neural Network Society's journal page (inns.org/neural-networks-journal, accessed October 2026).

The Society Journals Ecosystem

Neural Networks occupies a distinctive position in scholarly publishing: rather than belonging to a single publisher's portfolio, it is the archival journal of the world's three oldest neural-modelling societies — INNS, ENNS, and JNNS (with Elsevier's description adding APNNS). This society ownership gives the journal unusual stability: its readership is anchored by society members, all of whom receive an online subscription as part of membership (source: inns.org/neural-networks-journal).

The societies reinforce the journal's prestige through the annual Neural Networks Best Paper Award, recognising an outstanding paper published in the journal each year, presented at the International Joint Conference on Neural Networks — IJCNN (source: inns.org/neural-networks-journal).

INNS members receive an online subscription to the journal as part of membership, and the combined 50% open-access APC discount noted above. For prospective authors, this society ecosystem matters practically: members gain both read access and publishing support, and the journal's calls for papers and special issues — such as the open call on human-centred signals and multimodal behavioural data with a December 2026 deadline — are announced on the journal's ScienceDirect page (sources: inns.org; sciencedirect.com/journal/neural-networks).

Key Facts

Frequently Asked Questions

What is Neural Networks journal?

Neural Networks is a monthly peer-reviewed scientific journal published by Elsevier and the archival journal of the International Neural Network Society (INNS), the European Neural Network Society (ENNS), and the Japanese Neural Network Society (JNNS). It publishes research on all aspects of neural networks, including deep learning, from cognitive modeling and computational neuroscience to engineering applications (sources: sciencedirect.com/journal/neural-networks; inns.org).

When was Neural Networks founded?

The journal was established in 1988, with founding editor-in-chief Stephen Grossberg of Boston University (source: the journal's Elsevier description as recorded in standard references).

Who publishes Neural Networks?

Elsevier publishes the journal (Elsevier Ltd / Elsevier BV); it is the archival journal of the INNS, ENNS, and JNNS neural-network societies, with a subscription included in each society's membership (sources: sciencedirect.com; inns.org).

Who is the Editor-in-Chief of Neural Networks?

Taro Toyoizumi of the RIKEN Center for Brain Science (Wako, Japan) is Editor-in-Chief, supported by co-editors-in-chief Mauro Forti (University of Siena), Robert Kozma (University of Memphis), and Tongliang Liu (University of Sydney) (source: inns.org/neural-networks-journal, accessed October 2026).

What is the scope of Neural Networks?

The journal covers the full range of neural-network research — cognitive modeling, computational neuroscience, deep-learning algorithms, mathematical analyses, and engineering/technological applications — organised into five sections: Learning Systems, Cognitive Science, Neuroscience, Mathematical and Computational Analysis, and Engineering and Applications. Neural networks must be central to submissions (source: sciencedirect.com/journal/neural-networks/publish/guide-for-authors).

How does peer review work at Neural Networks?

The journal is peer-reviewed. For the specifics of the review process — review model, reviewer numbers, and editorial policies — consult the current Guide for Authors on the journal's ScienceDirect page before submitting.

How long does a decision take?

Elsevier's journal-insight statistics report a median of 9 days from submission to first decision, 78 days to a decision after review, 208 days to acceptance, and 8 days from acceptance to online publication (source: sciencedirect.com/journal/neural-networks, accessed October 2026).

What is the acceptance rate of Neural Networks?

The journal does not publicly disclose an acceptance rate. Figures on third-party websites are unofficial and should be treated with caution.

What is the APC for open access publishing?

The journal's ScienceDirect page lists an open access APC of USD 2,950 (excluding taxes), with possible reductions for authors qualifying for applicable discounts or institutional agreements. APCs change over time, so verify the current figure on ScienceDirect before submitting. INNS members receive a combined 50% discount (25% from the journal and 25% from INNS) on open-access article processing charges (sources: sciencedirect.com/journal/neural-networks; inns.org).

What article types does the journal publish?

Full-length Articles, Letters, Reviews and Tutorials, Letters to the Editor, editorials, and current events — organised across the five sections: Learning Systems, Cognitive Science, Neuroscience, Mathematical and Computational Analysis, and Engineering and Applications (per the journal's official description).

Where can I submit to Neural Networks?

Manuscripts are submitted electronically through Elsevier's editorial system, linked from the journal's ScienceDirect page and the Guide for Authors at sciencedirect.com/journal/neural-networks/publish/guide-for-authors.

Profile compiled October 2026 from official Elsevier/ScienceDirect and International Neural Network Society sources listed below. Metrics, APCs, and editorial personnel change over time — always verify current figures on the journal's official pages before citing or acting on them.