Deep Learning for Estimating Synaptic Health of Primary Neuronal Cell Culture

Alejo Nevado-Holgado; Andrey Kormilitzin; Caroline Woffindale; Elena Ribe; Francesca Nicholls; Noel Buckley; William H. Stone; Xinyu Yang

REVIEW 3 major objections 5 minor 46 references

Deep Learning for Estimating Synaptic Health of Primary Neuronal Cell Culture

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A fine-tuned residual CNN classifies raw cytoskeleton-channel images of rodent neurons as untreated or exposed to 30 µM Aβ(25–35) with 99.58% validation accuracy, then screens 36 compounds and finds none substantially protective.

desk verdict A useful screening application undermined by a train/test arithmetic inconsistency and an unvalidated proxy assumption; the 99.58% accuracy is not credible as reported. read the letter →

arxiv 1908.11399 v1 pith:77M3WRQB submitted 2019-08-29 eess.IV cs.LGq-bio.QMstat.ML

Andrey Kormilitzin , Xinyu Yang , William H. Stone , Caroline Woffindale , Francesca Nicholls , Elena Ribe , Alejo Nevado-Holgado , Noel Buckley This is my paper · ORCID

classification eess.IVcs.LGq-bio.QMstat.ML

keywords deeplearningconvolutionalneuralnetworkresidualconnectionstransferAlzheimer'sdiseaseamyloid-betasynaptichealthhigh-throughputscreening

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to replace the laborious feature-extraction route to assessing neuronal health in high-throughput drug screens with a single convolutional network that reads raw microscope images. It claims a fine-tuned residual CNN distinguishes untreated rodent primary neurons from neurons exposed to 30 µM of the amyloid-β fragment Aβ(25–35), reaching 99.58% validation accuracy on the cytoskeleton (Cy5) channel alone. Applied to 36 candidate compounds at three doses, the model reported no substantial protective effect against Aβ-induced synaptic loss, a result the authors say matches a feature-extraction-based statistical screen. The value of the claim, if true, is that synapse-loss detection becomes fast and feature-free, allowing large-scale screens that are impractical with hand-crafted image statistics.

What carries the argument

The carrying object is a fine-tuned ResNet18, a residual convolutional neural network with skip connections, pre-trained on a large natural-image corpus and re-trained on 2048×2048 Cy5 fluorescence images of neuronal cytoskeleton. Transfer learning, data augmentation, and dropout let the network learn features directly from pixels; Grad-CAM then maps which pixels drive each decision, showing attention on neurites. In screening, the network's per-image output is averaged over field views and wells to give a treatment-level score used to assign each condition to the untreated or Aβ-treated class.

What would settle it

Run the trained model on images from a plate containing a compound with independently confirmed synaptic rescue, for example by PSD-95 puncta counts; if those wells still receive Aβ-treated scores while the independent measure shows restored synapses, the screening criterion fails. Alternatively, compute validation accuracy separately for every held-out plate; if any plate's accuracy falls near chance, the 99.58% figure does not generalize across plates.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that the morphological signature of Aβ(25–35) toxicity in primary neuronal culture is learnable from raw Cy5-channel pixels: a ResNet18 initialized with pretrained weights and fine-tuned on 6,480 untreated and 6,480 treated images separates the two conditions with 99.58% validation accuracy. The same model, with predictions averaged over field views and wells, classifies cultures treated with compound plus Aβ as Aβ-treated unless the compound restores an untreated-like morphology. Using that criterion, none of the 36 screened compounds at 1, 3, or 10 µM showed substantial protection, and this negative result was corroborated by the standard feature-extraction and statistical-testing pipeline.

Load-bearing premise

The screen can only detect protection that makes an Aβ-exposed culture look like an untreated one; a compound that preserves synapses through a different visible morphology would be scored as ineffective.

Editorial extensions

If this is right

A high-throughput screen can be run on raw images: about 60 seconds per plate in inference versus roughly 15 hours for the traditional feature-extraction pipeline on 48 CPUs, the paper reports.
None of 36 candidate compounds at three doses substantially protected against Aβ(25–35)-induced synaptic loss under this assay's conditions.
The binary model's score can flag ambiguous wells for manual inspection, since most wells score near 0 or 1 while contaminated or blurry wells fall in between.
The approach is designed to transfer to other assays and, ultimately, human cells, since only raw pixels are needed and pretrained weights are reused.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Because the classifier is trained only on vehicle- and Aβ-treated endpoints, its screening utility depends on protection 'looking like' the untreated class; a compound that protects synapses through a different morphology would be scored as non-protective even if biologically effective.
A quantitative extension the paper does not report would correlate the model's continuous prediction score with independent synapse-density measures, such as PSD-95 puncta counts, across doses, turning the screen from binary to graded.
The single 34/2 plate split could be stress-tested by per-plate cross-validation; if accuracy varies sharply across plates, the 99.58% figure is a property of the chosen plates rather than the assay generally.
Feeding all four stains, including nuclear, pre-synaptic, and post-synaptic channels, into a multi-channel model could reveal whether the cytoskeleton channel alone carries the full signal or whether the other stains add independent predictive information.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Two things you should know. First, this is a legitimate new application: a stock ResNet18 distinguishes untreated neurons from Aβ(25-35)-treated neurons in raw Cy5 images, and the authors use it to screen 36 compounds, reporting no protective effect. The speed claim (60 seconds per plate vs 15 hours with CellProfiler) is a real practical point. Second, the headline 99.58% accuracy is not established. The text says 34 of 36 plates were used for training and 2 for testing, but the stated training set of 6,480 images per class equals all 36 plates (36 × 6 wells × 30 images = 6,480), not 34 plates (34 × 6 × 30 = 6,120). Either the held-out plates leaked into training or the validation number was computed with overlap. The Grad-CAM figures come from those same two plates, so they don't independently certify out-of-sample behavior. No code or data are provided to audit the split.

What the paper does well: the assay design is clearly described, the binary task is well motivated, and the screening outcome was cross-checked with CellProfiler, at least in the text. The compute comparison is worth reporting. This is a modest but real proof-of-concept for deep-learning-based screening in neurodegeneration.

The soft spots, in order. The split inconsistency is the load-bearing one; it undermines the accuracy claim and therefore the screening interpretation. Second, the screening logic assumes that a protective compound will push the morphology of Aβ-treated cells toward the untreated class. That is a reasonable first-pass hypothesis but it is unvalidated. The classifier sees only two extreme conditions; partial protection or alternative mechanisms could produce the same 'no protection' score. The CellProfiler confirmation is mentioned but no data are shown, so it can't be checked. Minor: the closing paragraph on signature methods is unrelated to the experiments and reads like an advertisement; it should be cut. These issues are fixable: report exact per-class counts, release code and data (or at least the split and inference code), and validate the proxy against a known protective compound or against the CellProfiler features.

Who should read this: people building high-content screening pipelines for Alzheimer's drug discovery. It deserves a serious referee because the application is relevant and the problems are addressable. As submitted, I would not accept it; I would ask for clarification of the split, the data/code, and a direct validation of the protective-effect proxy.

Referee Report

3 major / 5 minor

Summary. The paper presents a binary image classifier, based on a fine-tuned ResNet18, that distinguishes vehicle-control primary rodent neuronal cultures from cultures treated with 30 µM of Aβ(25-35), using raw Cy5-channel microscopy images. The authors report a validation accuracy of 99.58% and then apply the classifier to screen 36 candidate compounds for protective effects against Aβ-induced synaptic loss, concluding that none shows a substantial protective effect. The paper also includes Grad-CAM visualizations as a sanity check that the model attends to neurite-like structures and reports a computational speed advantage over a CellProfiler-based pipeline.

Significance. If the accuracy estimate were reliable, the claimed result would be practically useful: a raw-pixel CNN classifier that is as accurate as claimed and runs in about 60 seconds per plate could substantially accelerate high-throughput screening of compounds for morphological neuroprotection. The paper also deserves credit for applying Grad-CAM to check that decisions are not driven by obvious artifacts, and for being explicit about the screening logic. However, the central quantitative claim is undermined by an arithmetic inconsistency in the training/test split, and the screening conclusion rests on a strong interpretive assumption that is not validated. As written, the 99.58% figure is not an established out-of-sample estimate, and the paper's main conclusions therefore require revision rather than being directly acceptable.

major comments (3)

[§3.1, §3.2, and §4] The reported validation accuracy is not established as an out-of-sample estimate because the stated training-set size is inconsistent with the described plate split. Section 3.1 states that each well is imaged 30 times, and Table 1 gives six vehicle-control wells and six Aβ-only wells per plate; Section 4 states that 34 of 36 plates were used for training. That yields 34 × 6 × 30 = 6,120 images per class, not the 6,480 per class reported in Section 3.2. The stated 6,480 corresponds to using all 36 plates. Therefore either the two 'held-out' plates were included in fine-tuning, or the validation numbers in Table 2 were computed with training data leakage. In either case, the 99.58% validation accuracy cannot be interpreted as an independent test of generalization, and no code or data are provided to audit the split.
[§3.2 and §5] The screening conclusion does not follow from the classification accuracy alone. Section 3.2 states that the trained model was used to determine whether cells treated with a compound plus Aβ(25-35) are 'more similar' to cells treated with Aβ(25-35) alone, and Section 5 concludes that none of the 36 compounds has a substantial protective effect. Because the classifier was trained only on the two extreme conditions, it cannot distinguish partial protection, protective mechanisms that do not restore the specific morphology of the untreated class, or off-target effects that change image features for unrelated reasons. The conclusion is a domain assumption about what protection should look like in the learned feature space, and the paper provides no independent validation of that assumption.
[§5] The claim that the screening result was 'confirmed' by a CellProfiler-based statistical pipeline is not supported by any reported analysis. Section 5 states this confirmation, but the paper does not give the CellProfiler feature measurements, the statistical test results, or a comparison table for the 36 compounds. As written, this confirmation is unverifiable and cannot be used to strengthen the main screening conclusion.

minor comments (5)

[§3.2] There is a typo: 'the the model was applied to screen candidate compounds' should read 'the model was applied to screen candidate compounds.'
[Captions of Figures 2 and 3] The figure captions describe the second component of each triple as 'Aβ(25−30)', but the text and methods consistently use 'Aβ(25−35)'; the captions should be corrected for consistency.
[§5] The word 're-suing' in the last paragraph should be 're-using.'
[§4 and Appendix A] The reference to 'Table 5 in A' should refer to 'Appendix A,' and Table 5 should be placed in the appendix rather than left dangling in the results section.
[§3.2] The model name is written inconsistently as 'Resnet18' in the text and 'ResNet18' in the abstract; please standardize the capitalization.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the 99.58% accuracy is an externally benchmarked classification result and the screening conclusion is an explicitly stated biological assumption, not a construction that renames the model's inputs.

full rationale

The paper's central accuracy claim (Table 2: validation accuracy 99.58%) is measured against independent treatment labels ('untreated' vehicle control vs. 30 µM Aβ(25-35)) on images that were not used for training (two plates held out, per Section 4). This is an external benchmark and not a fitted input masquerading as a prediction. The screening step is the only place where a conclusion could in principle be generated by the classifier's own labels: Section 3.2 states that once trained, the model was applied 'to determine whether images of cells treated with a compound + Aβ(25-35) were more similar to those treated with Aβ(25-35) alone.' That is a substantive biological assumption (protection implies untreated-like morphology), not a circular derivation; the conclusion is not guaranteed by the model's construction, and the paper separately claims confirmation via CellProfiler. Self-citations in the manuscript (e.g., refs. 34, 38, 39) concern the unrelated signature-method future-work discussion and do not support the classification or screening claims. The arithmetic discrepancy noted in the skeptical summary (34 plates × 6 wells × 30 field views = 6,120 images per class, not 6,480) is a data-integrity/validation concern outside the circularity rubric, since it does not show that the accuracy was defined into existence. Accordingly, no circular step meeting the evidentiary bar is identified.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The study introduces no new entities. It relies on standard deep learning assumptions and a domain-specific proxy for protection, plus hand-chosen hyperparameters.

free parameters (4)

Decision threshold = 0.5
Used to convert predicted scores to binary classes; the screening conclusions depend on this arbitrary cutoff.
Fine-tuning learning rates = 1e-3 then 1e-4
Chosen by hand, no sensitivity analysis reported.
Number of training epochs = 10 + 10
Chosen by hand; no early stopping or validation-based selection is described.
Dropout rate = 0.35
Chosen by hand to 'lead to stable convergence', no ablation.

assumptions (5)

domain assumption ImageNet-pretrained weights transfer to fluorescence microscopy images.
ResNet18 initialized with ImageNet weights; no ablation shows this is necessary or optimal for this domain.
domain assumption The Cy5 channel alone captures the morphological changes induced by Aβ(25-35).
The model uses only the Cy5 stain; other channels (DAPI, dsRed, FITC) are collected but not used, with no validation that they are redundant.
domain assumption The two held-out assay plates are representative of plate-to-plate variability.
Validation is performed on only 2 of 36 plates; no explicit normalization for batch effects or plate-specific staining artifacts is reported.
domain assumption A protective compound shifts cell morphology toward the untreated class.
Core screening assumption discussed in Section 3.2; the binary classifier cannot directly measure protection.
standard math Standard stochastic gradient descent training converges to a generalizing solution.
No formal guarantee, but standard practice for CNNs.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Deep Learning for Estimating Synaptic Health of Primary Neuronal Cell Culture." pith.science (2026). https://pith.science/paper/77M3WRQB

@misc{pith2026190811399,
  author       = {Pith},
  title        = {Pith review of: Deep Learning for Estimating Synaptic Health of Primary Neuronal Cell Culture},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/77M3WRQB}},
  note         = {Machine review of arXiv:1908.11399}
}

abstract

Understanding the morphological changes of primary neuronal cells induced by chemical compounds is essential for drug discovery. Using the data from a single high-throughput imaging assay, a classification model for predicting the biological activity of candidate compounds was introduced. The image recognition model which is based on deep convolutional neural network (CNN) architecture with residual connections achieved accuracy of 99.6$\%$ on a binary classification task of distinguishing untreated and treated rodent primary neuronal cells with Amyloid-$\beta_{(25-35)}$.

Figures

Figures reproduced from arXiv: 1908.11399 by the authors.

**Figure 1.** An assay plate with wells. Well position index Compound Aβ B{2,3,4}, C{5,6,7} 0 0 C{2,3,4}, D{5,6,7} 0 30 D{2,3,4}, E{5,6,7} 1 0 E{2,3,4}, F{5,6,7} 1 30 F{2,3,4}, G{5,6,7} 3 0 G{2,3,4}, H{5,6,7} 3 30 H{2,3,4}, J{5,6,7} 10 0 F{2,3,4}, B{5,6,7} 10 30 [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗

**Figure 2.** Random images from the plate with Raubasine compound. Each image is described by a triple (a, b, c), where a - compound dose (µM), b - Aβ(25−30) dose (µM) and the predicted score from the model that the image was classified correctly. Left column corresponds to treatments with only Raubasine at various regimes and without added 30 µM of Aβ(25−30). 8 [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗

**Figure 3.** Random images from the plate with Thiamphenicol compound. Each image is described by a triple (a, b, c), where a - compound dose (µM), b - Aβ(25−30) dose (µM) and the predicted score from the model that the image was classified correctly. Left column corresponds to treatments with only Thiamphenicol at various regimes and without added 30 µM of Aβ(25−30). 9 [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: A distribution of the predicted scores for [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]

**Figure 5.** Figure 5: A distribution of the predicted scores for [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

46 extracted references · 28 canonical work pages

[1]

Contribution of alzheimer disease to mortality in the united states

Bryan D James, Sue E Leurgans, Liesi E Hebert, Paul A Scherr, Kristine Yaffe, and David A Bennett. Contribution of alzheimer disease to mortality in the united states. Neurology, 82(12):1045–1050, 2014

work page 2014
[2]

Uk research spend in 2008 and 2012: comparing stroke, cancer, coronary heart disease and dementia

Ramon Luengo-Fernandez, Jose Leal, and Alastair Gray. Uk research spend in 2008 and 2012: comparing stroke, cancer, coronary heart disease and dementia. BMJ open, 5(4):e006648, 2015

work page 2008
[3]

Guy M McKhann, David S Knopman, Howard Chertkow, Bradley T Hyman, Clifford R Jack Jr, Claudia H Kawas, William E Klunk, Walter J Koroshetz, Jennifer J Manly, Richard Mayeux, et al. The diagnosis of dementia due to alzheimer’s disease: Recommendations from the national institute on aging-alzheimer’s association workgroups on diagnostic guidelines for alzhe...

work page 2011
[4]

The amyloid beta peptide: a chemist’s perspective

Ian W Hamley. The amyloid beta peptide: a chemist’s perspective. role in alzheimer’s and ﬁbrillization. Chemical reviews, 112(10):5147–5192, 2012

work page 2012
[5]

Discovery, replication and extension study of plasma proteomic biomarkers relating to brain amyloid burden and alzheimer’s disease progression

Sarah Westwood, Alison L Baird, Sneha N Anand, Liu Shi, Alejo J Nevado-Holgado, Andrey Kormilitzin, Abdul Hye, Nicholas J Ashton, Angharad Morgan, Samuel Touchard, et al. Discovery, replication and extension study of plasma proteomic biomarkers relating to brain amyloid burden and alzheimer’s disease progression. Alzheimer’s & Dementia: The Journal of the...

work page 2018
[6]

Overview of deep learning in medical imaging

Kenji Suzuki. Overview of deep learning in medical imaging. Radiological physics and technology, 10(3):257–273, 2017

work page 2017
[7]

Detecting cancer metastases on gigapixel pathology images

Yun Liu, Krishna Gadepalli, Mohammad Norouzi, George E Dahl, Timo Kohlberger, Aleksey Boyko, Subhashini Venugopalan, Aleksei Timofeev, Philip Q Nelson, Greg S Corrado, et al. Detecting cancer metastases on gigapixel pathology images. arXiv preprint arXiv:1703.02442, 2017

arXiv 2017
[8]

Artiﬁcial intelligence–based breast cancer nodal metastasis detection: Insights into the black box for pathologists

Yun Liu, Timo Kohlberger, Mohammad Norouzi, George E Dahl, Jenny L Smith, Arash Mohtashamian, Niels Olson, Lily H Peng, Jason D Hipp, and Martin C Stumpe. Artiﬁcial intelligence–based breast cancer nodal metastasis detection: Insights into the black box for pathologists. Archives of Pathology & Laboratory Medicine

work page

Show all 46 references

[9]

Deep learning in image cytometry: a review

Anindya Gupta, Philip J Harrison, Håkan Wieslander, Nicolas Pielawski, Kimmo Kartasalo, Gabriele Partel, Leslie Solorzano, Amit Suveer, Anna H Klemm, Ola Spjuth, et al. Deep learning in image cytometry: a review. Cytometry Part A, 95(4):366–380, 2019

2019
[10]

Machine learning methods for histopathological image analysis

Daisuke Komura and Shumpei Ishikawa. Machine learning methods for histopathological image analysis. Computational and structural biotechnology journal , 16:34–42, 2018

2018
[11]

Deep learning-based virtual histology staining using auto-ﬂuorescence of label-free tissue

Yair Rivenson, Hongda Wang, Zhensong Wei, Yibo Zhang, Harun Gunaydin, and Aydogan Ozcan. Deep learning-based virtual histology staining using auto-ﬂuorescence of label-free tissue. arXiv preprint arXiv:1803.11293, 2018

2018 arXiv
[12]

Virtual histological staining of unlabelled tissue-autoﬂuorescence images via deep learning

Yair Rivenson, Hongda Wang, Zhensong Wei, Kevin de Haan, Yibo Zhang, Yichen Wu, Harun Günaydın, Jonathan E Zuckerman, Thomas Chong, Anthony E Sisk, et al. Virtual histological staining of unlabelled tissue-autoﬂuorescence images via deep learning. Nature biomedical engineering...

2019
[13]

Deep learning enables cross-modality super-resolution in ﬂuorescence microscopy

Hongda Wang, Yair Rivenson, Yiyin Jin, Zhensong Wei, Ronald Gao, Harun Günaydın, Laurent A Bentolila, Comert Kural, and Aydogan Ozcan. Deep learning enables cross-modality super-resolution in ﬂuorescence microscopy. Nat. Methods, 16:103–110, 2019

2019
[14]

Reconstructing cell cycle and disease progression using deep learning

Philipp Eulenberg, Niklas Köhler, Thomas Blasi, Andrew Filby, Anne E Carpenter, Paul Rees, Fabian J Theis, and F Alexander Wolf. Reconstructing cell cycle and disease progression using deep learning. Nature communications, 8(1):463, 2017

2017
[15]

Repurposing high-throughput image assays enables biological activity prediction for drug discovery

Jaak Simm, Günter Klambauer, Adam Arany, Marvin Steijaert, Jörg Kurt Wegner, Emmanuel Gustin, Vladimir Chupakhin, Yolanda T Chong, Jorge Vialard, Peter Buijnsters, et al. Repurposing high-throughput image assays enables biological activity prediction for drug discovery. Cell c...

2018
[16]

Cellproﬁler: image analysis software for identifying and quantifying cell phenotypes

Anne E Carpenter, Thouis R Jones, Michael R Lamprecht, Colin Clarke, In Han Kang, Ola Friman, David A Guertin, Joo Han Chang, Robert A Lindquist, Jason Moffat, et al. Cellproﬁler: image analysis software for identifying and quantifying cell phenotypes. Genome biology, 7(10):R100, 2006

2006
[17]

Deep learning

Yann LeCun, Yoshua Bengio, and Geoffrey Hinton. Deep learning. nature, 521(7553):436, 2015

2015
[18]

Deep residual learning for image recognition

Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 770–778, 2016

2016
[19]

J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei. ImageNet: A Large-Scale Hierarchical Image Database. In CVPR09, 2009

2009
[20]

Automatic differentiation in pytorch

Adam Paszke, Sam Gross, Soumith Chintala, Gregory Chanan, Edward Yang, Zachary DeVito, Zeming Lin, Alban Desmaison, Luca Antiga, and Adam Lerer. Automatic differentiation in pytorch. 2017

2017
[21]

Jeremy Howard et al. fastai. https://github.com/fastai/fastai, 2018

2018
[22]

Improving neural networks by preventing co-adaptation of feature detectors

Geoffrey E Hinton, Nitish Srivastava, Alex Krizhevsky, Ilya Sutskever, and Ruslan R Salakhutdinov. Improving neural networks by preventing co-adaptation of feature detectors. arXiv preprint arXiv:1207.0580, 2012

2012 arXiv
[23]

A disciplined approach to neural network hyper-parameters: Part 1–learning rate, batch size, momentum, and weight decay

Leslie N Smith. A disciplined approach to neural network hyper-parameters: Part 1–learning rate, batch size, momentum, and weight decay. arXiv preprint arXiv:1803.09820, 2018

2018 arXiv
[24]

Universal language model ﬁne-tuning for text classiﬁcation

Jeremy Howard and Sebastian Ruder. Universal language model ﬁne-tuning for text classiﬁcation. arXiv preprint arXiv:1801.06146, 2018

2018 arXiv
[25]

Visualizing the loss landscape of neural nets

Hao Li, Zheng Xu, Gavin Taylor, Christoph Studer, and Tom Goldstein. Visualizing the loss landscape of neural nets. In Advances in Neural Information Processing Systems , pages 6389–6399, 2018

2018
[26]

Searching for activation functions

Prajit Ramachandran, Barret Zoph, and Quoc V Le. Searching for activation functions. arXiv preprint arXiv:1710.05941, 2017

2017 arXiv
[27]

Grad-cam: Visual explanations from deep networks via gradient- based localization

Ramprasaath R Selvaraju, Michael Cogswell, Abhishek Das, Ramakrishna Vedantam, Devi Parikh, Dhruv Batra, et al. Grad-cam: Visual explanations from deep networks via gradient- based localization. In ICCV, pages 618–626, 2017

2017
[28]

Aβ inﬂuences cytoskeletal signaling cascades with consequences to alzheimer’s disease

Ana Gabriela Henriques, Joana Machado Oliveira, Liliana Patrícia Carvalho, and Odete AB da Cruz e Silva. Aβ inﬂuences cytoskeletal signaling cascades with consequences to alzheimer’s disease. Molecular neurobiology, 52(3):1391–1407, 2015

2015
[29]

Alexander B. Jung. imgaug. https://github.com/aleju/imgaug, 2018. [Online; accessed 25-Aug-2019]

2018
[30]

Segem: efﬁcient image analysis for high-resolution connectomics

Manuel Berning, Kevin M Boergens, and Moritz Helmstaedter. Segem: efﬁcient image analysis for high-resolution connectomics. Neuron, 87(6):1193–1206, 2015

2015
[31]

Analyzing image segmentation for connectomics

Stephen M Plaza and Jan Funke. Analyzing image segmentation for connectomics. Frontiers in Neural Circuits, 12, 2018

2018
[32]

Synapse-aware skeleton generation for neural circuits

Brian Matejek, Donglai Wei, Xueying Wang, Jinglin Zhao, Kálmán Palágyi, and Hanspeter Pﬁster. Synapse-aware skeleton generation for neural circuits
[33]

Rough paths, signatures and the modelling of functions on streams

Terry Lyons. Rough paths, signatures and the modelling of functions on streams. arXiv preprint arXiv:1405.4537, 2014. 6

2014 arXiv
[34]

A primer on the signature method in machine learning

Ilya Chevyrev and Andrey Kormilitzin. A primer on the signature method in machine learning. arXiv preprint arXiv:1603.03788, 2016

2016
[35]

Sparse arrays of signatures for online character recognition

Benjamin Graham. Sparse arrays of signatures for online character recognition. arXiv preprint arXiv:1308.0371, 2013

2013 arXiv
[36]

Learning spatial- semantic context with fully convolutional recurrent network for online handwritten chinese text recognition

Zecheng Xie, Zenghui Sun, Lianwen Jin, Hao Ni, and Terry Lyons. Learning spatial- semantic context with fully convolutional recurrent network for online handwritten chinese text recognition. IEEE transactions on pattern analysis and machine intelligence , 40(8):1903–1917, 2017

1903
[37]

Leveraging the path signature for skeleton-based human action recognition

Weixin Yang, Terry Lyons, Hao Ni, Cordelia Schmid, Lianwen Jin, and Jiawei Chang. Leveraging the path signature for skeleton-based human action recognition. arXiv preprint arXiv:1707.03993, 2017

2017 arXiv
[38]

Application of the signature method to pattern recognition in the cequel clinical trial

AB Kormilitzin, KEA Saunders, PJ Harrison, JR Geddes, and TJ Lyons. Application of the signature method to pattern recognition in the cequel clinical trial. arXiv preprint arXiv:1606.02074, 2016

2016 arXiv
[39]

Detecting early signs of depressive and manic episodes in patients with bipolar disorder using the signature-based model

Andrey Kormilitzin, Kate EA Saunders, Paul J Harrison, John R Geddes, and Terry Lyons. Detecting early signs of depressive and manic episodes in patients with bipolar disorder using the signature-based model. arXiv preprint arXiv:1708.01206, 2017

2017 arXiv
[40]

A signature-based machine learning model for bipolar disorder and borderline personality disorder

Imanol Perez Arribas, Kate Saunders, Guy Goodwin, and Terry Lyons. A signature-based machine learning model for bipolar disorder and borderline personality disorder. arXiv preprint arXiv:1707.07124, 2017

2017 arXiv
[41]

Using path signatures to predict a diagnosis of alzheimer’s disease

PJ Moore, J Gallacher, and TJ Lyons. Using path signatures to predict a diagnosis of alzheimer’s disease. arXiv preprint arXiv:1808.05865, 2018

2018 arXiv
[42]

Random forest prediction of alzheimer’s disease using pairwise selection from time series data

PJ Moore, TJ Lyons, John Gallacher, Alzheimer’s Disease Neuroimaging Initiative, et al. Random forest prediction of alzheimer’s disease using pairwise selection from time series data. PloS one, 14(2):e0211558, 2019

2019
[43]

Derivatives pricing using signature payoffs

Imanol Perez Arribas. Derivatives pricing using signature payoffs. arXiv preprint arXiv:1809.09466, 2018

2018 arXiv
[44]

Extracting information from the signature of a ﬁnancial data stream

Lajos Gergely Gyurkó, Terry Lyons, Mark Kontkowski, and Jonathan Field. Extracting information from the signature of a ﬁnancial data stream. arXiv preprint arXiv:1307.7244 , 2013

2013 arXiv
[45]

Learning from the past, predicting the statistics for the future, learning an evolving system

Daniel Levin, Terry Lyons, and Hao Ni. Learning from the past, predicting the statistics for the future, learning an evolving system. arXiv preprint arXiv:1309.0260, 2013

2013 arXiv
[46]

Deep signatures

Patric Bonnier, Patrick Kidger, Imanol Perez Arribas, Cristopher Salvi, and Terry Lyons. Deep signatures. arXiv preprint arXiv:1905.08494, 2019. A List of compounds used for training and screening Amprolium Bucladesine Camptothecin Carbimazol Cefapirin Chloropyrazine Chloroqui...

1905 arXiv

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

Contribution of alzheimer disease to mortality in the united states

Bryan D James, Sue E Leurgans, Liesi E Hebert, Paul A Scherr, Kristine Yaffe, and David A Bennett. Contribution of alzheimer disease to mortality in the united states. Neurology, 82(12):1045–1050, 2014

work page 2014

[2] [2]

Uk research spend in 2008 and 2012: comparing stroke, cancer, coronary heart disease and dementia

Ramon Luengo-Fernandez, Jose Leal, and Alastair Gray. Uk research spend in 2008 and 2012: comparing stroke, cancer, coronary heart disease and dementia. BMJ open, 5(4):e006648, 2015

work page 2008

[3] [3]

Guy M McKhann, David S Knopman, Howard Chertkow, Bradley T Hyman, Clifford R Jack Jr, Claudia H Kawas, William E Klunk, Walter J Koroshetz, Jennifer J Manly, Richard Mayeux, et al. The diagnosis of dementia due to alzheimer’s disease: Recommendations from the national institute on aging-alzheimer’s association workgroups on diagnostic guidelines for alzhe...

work page 2011

[4] [4]

The amyloid beta peptide: a chemist’s perspective

Ian W Hamley. The amyloid beta peptide: a chemist’s perspective. role in alzheimer’s and ﬁbrillization. Chemical reviews, 112(10):5147–5192, 2012

work page 2012

[5] [5]

Discovery, replication and extension study of plasma proteomic biomarkers relating to brain amyloid burden and alzheimer’s disease progression

Sarah Westwood, Alison L Baird, Sneha N Anand, Liu Shi, Alejo J Nevado-Holgado, Andrey Kormilitzin, Abdul Hye, Nicholas J Ashton, Angharad Morgan, Samuel Touchard, et al. Discovery, replication and extension study of plasma proteomic biomarkers relating to brain amyloid burden and alzheimer’s disease progression. Alzheimer’s & Dementia: The Journal of the...

work page 2018

[6] [6]

Overview of deep learning in medical imaging

Kenji Suzuki. Overview of deep learning in medical imaging. Radiological physics and technology, 10(3):257–273, 2017

work page 2017

[7] [7]

Detecting cancer metastases on gigapixel pathology images

Yun Liu, Krishna Gadepalli, Mohammad Norouzi, George E Dahl, Timo Kohlberger, Aleksey Boyko, Subhashini Venugopalan, Aleksei Timofeev, Philip Q Nelson, Greg S Corrado, et al. Detecting cancer metastases on gigapixel pathology images. arXiv preprint arXiv:1703.02442, 2017

arXiv 2017

[8] [8]

Artiﬁcial intelligence–based breast cancer nodal metastasis detection: Insights into the black box for pathologists

Yun Liu, Timo Kohlberger, Mohammad Norouzi, George E Dahl, Jenny L Smith, Arash Mohtashamian, Niels Olson, Lily H Peng, Jason D Hipp, and Martin C Stumpe. Artiﬁcial intelligence–based breast cancer nodal metastasis detection: Insights into the black box for pathologists. Archives of Pathology & Laboratory Medicine

work page