NeuroPlast: a learnable activation function evaluated under knowledge distillation for medical image classification.

Overview

1. Department of Electrical Engineering, École de technologie supérieure, Montréal, QC, Canada
2. Department of Acute Medicine, Hôpital Universitaire de Genève, Genève, Switzerland
3. Institut de recherche Robert-Sauvé en santé et en sécurité du travail (IRSST), Montréal, QC, Canada

Abstract

Deep neural networks for medical image classification rely almost exclusively on fixed activation functions such as ReLU. We introduce NeuroPlast, a parametric activation function whose four differentiable components (a shifted sigmoid modeling NMDA-type voltage gating, a Gaussian plateau inspired by AMPA receptor scaling, an excitatory rectifier, and an inhibitory leak) are combined through six learnable parameters. Two mixing strategies are evaluated: a static variant with fixed learned weights, and a metaplastic variant whose mixing coefficients adapt per sample via a lightweight squeeze-excite gate conditioned on channel statistics. NeuroPlast is embedded within NADN Ultra, a 13.5 M-parameter residual convolutional neural network (CNN) with convolutional block attention modules (CBAM), trained entirely from scratch through a two-phase knowledge distillation (KD) pipeline. The teacher is a fine-tuned EfficientNet-B0; the student combines logit-level KD with optional feature-level alignment losses. Across four medical imaging benchmarks (Brain Tumor MRI, 7,200 images, 4 classes; Chest X-ray Pneumonia, 5,856 images, 2 classes; Skin Cancer HAM10000, 10,015 images, 7 classes; COVID-19 Radiography, 10,848 images, 2 classes), evaluated under five-fold stratified cross-validation with 95% confidence intervals, the static-KD variant reaches 99.4–99.5% accuracy on COVID-19 X-ray across three seeds, on par with pretrained EfficientNet-B0 (99.18%, within replication noise) and above ResNet-18 (98.69%). It closes 65% of the accuracy gap on Brain Tumor MRI (98.29% vs. 99.13%), reaches 92.5% on Chest X-ray under a uniform class-balancing rule, and 79.5% on Skin Cancer under lesion-grouped cross-validation. The metaplastic variant achieves 99.28% on COVID-19 X-ray, still above pretrained baselines, but does not consistently outperform the static version, a negative result analyzed through ablation experiments. On three tabular medical datasets spanning three orders of magnitude in sample size (569 to 253,680), NeuroPlast matches five established activations within ±1.7 percentage points; at 253 K samples all activations converge within 0.09 pp. Our findings indicate that knowledge distillation is the primary enabler for from-scratch architectures to approach pretrained-level performance on medical images; under an identical distillation pipeline, NeuroPlast adds a small but consistent gain over ReLU, GELU, Swish, Mish, and PReLU, leading on all four imaging benchmarks by margins below one point.

Reproduced under the paper's license (CC BY), from the paper cited above.

Code

The paper links to its data, not to its authors' code: see the Data section.

Tracing map

A tracing map links a paper to the code its authors published: this paper has none, so it has no map.

Data

Datasets cited

• data.mendeley.com/datasets/rscbjbr9sj, at Mendeley Data; found in “Data availability statement”
• doi:10.7910/dvn/dbw86t, at the source; found in “Data availability statement”
• kaggle.com/datasets/masoudnickparvar, at Kaggle; found in “Data availability statement”
• kaggle.com/datasets/tawsifurrahman, at Kaggle; found in “Data availability statement”

Data availability statement

Publicly available datasets were analysed in this study: Brain Tumor MRI (https://www.kaggle.com/datasets/masoudnickparvar/brain-tumor-mri-dataset); Chest X-ray Pneumonia (Kermany et al., 2018; https://data.mendeley.com/datasets/rscbjbr9sj/2); Skin Cancer HAM10000 (Tschandl et al., 2018; https://doi.org/10.7910/DVN/DBW86T); COVID-19 Radiography (Chowdhury et al., 2020; https://www.kaggle.com/datasets/tawsifurrahman/covid19-radiography-database); Wisconsin Breast Cancer, Pima Indians Diabetes, and Heart Disease BRFSS 2022 from the UCI Machine Learning Repository/CDC BRFSS.

Reproduced under the paper's license (CC BY), from the paper cited above.

Versions

The history of this record: each version stored by the harvester or made by a correction of its authors or of the maintainers of its code, and what changed in its facts. The texts of the paper (its abstract, its availability statements) are not part of it; versions that changed only those are not listed.

Version 2, 28 September 2026

• Funding: added Nvidia; École de technologie supérieure; Natural Sciences and Engineering Research Council of Canada

Version 1, 27 September 2026: the first record

Recorded: type, language, journal, volume, pages, dates, 7 authors, 8 keywords, 37 references.

Cite

This paper

Vaussenat, F., Bhattacharya, A., Payette, J., Desmettre, T., Saidi, A., Gagnon, G., & Cloutier, S. G. (2026). NeuroPlast: a learnable activation function evaluated under knowledge distillation for medical image classification. Frontiers in artificial intelligence, 9, 1849571. https://doi.org/10.3389/frai.2026.1849571

Similar papers

The papers with a page that share the most with this one: the tools found in their code, their categories, datasets, cited references and authors, the rarest counting most.

[1] doi:10.3389/fonc.2026.1834400

BrainFusionNet: an attention-augmented deep convolutional framework with hybrid loss optimisation and test-time augmentation for multi-class brain tumour detection in magnetic resonance images.

Journal: Frontiers in oncology

In common: kaggle.com/datasets/masoudnickparvar, methods / tools, other condition, 4 references

[2] doi:10.1038/s41598-026-55179-4 [code]

Vision-Language Models for automated quality control: a benchmarking framework and comprehensive study.

Journal: Scientific reports

In common: DOI 10.7910/dvn/dbw86t, methods / tools, other condition, 1 reference

[3] doi:10.1038/s41598-026-51120-x [code]

A compact and interpretable multi-source framework for heterogeneous medical image classification.

Journal: Scientific reports

In common: data.mendeley.com/datasets/rscbjbr9sj, methods / tools, other condition, 1 reference

[4] doi:10.1038/s41598-026-52658-6

A multi-level attention CNN-transformer based framework for the detection of brain tumor using regional dual-score explainability.

Journal: Scientific reports

In common: kaggle.com/datasets/masoudnickparvar, methods / tools, other condition, 1 reference

[5] doi:10.3389/fonc.2026.1806663

A denoising autoencoder-enhanced ResNet 50 framework for explainable brain tumor classification using MRI.

Journal: Frontiers in oncology

In common: kaggle.com/datasets/masoudnickparvar, methods / tools, other condition, 1 reference

[6] doi:10.3389/fmed.2026.1905061

Performance trade-offs between dense prediction and sparse query mechanisms for brain tumor MRI detection: a comparative study of YOLOv8 and RT-DETR.

Journal: Frontiers in medicine

In common: kaggle.com/datasets/masoudnickparvar, other condition, 1 reference

[7] doi:10.1038/s41598-026-61891-y

A class-wise quantum relational calibration network for brain tumor diagnosis.

Journal: Scientific reports

In common: kaggle.com/datasets/masoudnickparvar, other condition, 1 reference

[8] doi:10.1002/acm2.70560 [code]

Hybrid CNN-GCN framework for brain tumor MRI classification: A graph-based approach to smart healthcare diagnostics.

Journal: Journal of applied clinical medical physics

In common: kaggle.com/datasets/masoudnickparvar, other condition, 1 reference

[9] doi:10.1117/1.jbo.31.8.086006

Noise-aware dynamic convolution for improved generalizability of retinal disease diagnosis using optical coherence tomography images.

Journal: Journal of biomedical optics

In common: data.mendeley.com/datasets/rscbjbr9sj

[10] doi:

NeuroTrustNet: a cost-effective multimodal ensemble framework for brain tumor classification under cross-dataset variability

Journal: Frontiers in artificial intelligence

In common: kaggle.com/datasets/masoudnickparvar, methods / tools, other condition

Contribute

The authors of this paper can claim it, correct its record and validate its tracing map, and the maintainers of its code (its owner, or a public member of its organization) correct what it says of their repository; anyone signed in can ask for its removal. Every request goes to OSCR's own machine, which answers it; your account page follows them.

Request its removal

To ask OSCR to remove this record, the copies of its authors' scripts or its tracing map, use the removal request page: signed in, you say who you are, what to remove and why, then review and confirm the request. Published rules decide every request (how).

Discussion: questions and error reports about this paper and its code, from signed-in readers and its authors. It opens with sign-in.

Reproductions: reports from readers who ran the authors' code: what they reproduced, with which environment, commit and data. It opens with sign-in.

Activity: what happens around this paper: new versions of its record, its map's validation, discussions and reproductions. It opens with sign-in.