Abstract
Purpose
Point localisation is a critical aspect of many interventional planning procedures, specifically representing anatomical regions of interest or landmarks as individual points. This could be seen as analogous to the problem of visual search in cognitive psychology, in which this search is performed either: bottom-up, constructing increasingly abstract and coarse-resolution features over the entire image; or top-down, using contextual cues from the entire image to refine the scope of the region being investigated. Traditional convolutional neural networks use the former, but it is not clear if this is optimal. This article is a preliminary investigation as to how this motivation affects 3D point localisation in neuro-interventional planning.
Methods
Two neuro-imaging datasets were collected: one for cortical point localisation for repetitive transcranial magnetic stimulation and the other for sub-cortical anatomy localisation for deep brain stimulation. Four different frameworks were developed using top-down versus bottom-up paradigms as well as representing points as co-ordinates or heatmaps. These networks were applied to point localisation for transcranial magnetic stimulation and subcortical anatomy localisation. These networks were evaluated using cross-validation and a varying number of training datasets to analyse their sensitivity to quantity of training data.
Results
Each network shows increasing performance as the amount of available training data increases, with the co-ordinate-based top-down network consistently outperforming the others. Specifically, the top-down architectures tend to outperform the bottom-up ones. An analysis of their memory consumption also encourages the top-down co-ordinate based architecture as it requires significantly less memory than either bottom-up architectures or those representing their predictions via heatmaps.
Conclusion
This paper is a preliminary foray into a fundamental aspect of machine learning architectural design: that of the top-down/bottom-up divide from cognitive psychology. Although there are additional considerations within the particular architectures investigated that could affect these results and the number of architectures investigated is limited, our results do indicate that the less commonly used top-down paradigm could lead to more efficient and effective architectures in the future.
Similar content being viewed by others
Data availability
Result tables for all experiments are available on request. However, we can not provide sensitive patient data.
Code availability
Our code will be available on request.
Notes
http://www.itksnap.org/pmwiki/pmwiki.php?n=Convert3D.Convert3D, The specific commands used were swapdim, -resample-mm, and -pad-to.
References
He K, Sun J (2015) Convolutional neural networks at constrained time cost. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 5353–5360
Qiao S, Lin Z, Zhang J, Yuille A L (2019) Neural rejuvenation: improving deep network training by enhancing computational resource utilization. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp. 61–71
Girshick R, Donahue J, Darrell T, Malik J (2014) Rich feature hierarchies for accurate object detection and semantic segmentation. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 580–587
Girshick R (2015) Fast r-cnn. In: Proceedings of the IEEE international conference on computer vision, pp. 1440–1448
Ren S, He K, Girshick R, Sun J (2015) Faster r-cnn: towards real-time object detection with region proposal networks. In: Advances in neural information processing systems, vol. 28, pp. 91–99
Sugimori H, Kawakami M (2019) Automatic detection of a standard line for brain magnetic resonance imaging using deep learning. In: Applied Sciences, vol. 9, 3849
Yang X, Tang WT, Tjio G, Yeo SY, Su Y (2020) Automatic detection of anatomical landmarks in brain MR scanning using multi-task deep neural networks. In: Neurocomputing, vol. 396, pp. 514–521
Gohel B, Kumar L, Shah D (2023) Deep learning-based automated localisation of anterior commissure and posterior commissure landmarks in 3d space from three-plane 2d mri localiser slices of the brain. In: Procedia Computer Science, vol. 218, pp. 1027–1032
Baxter JS, Bui QA, Maguet E, Croci S, Delmas A, Lefaucheur J-P, Bredoux L, Jannin P (2021) Automatic cortical target point localisation in MRI for transcranial magnetic stimulation via a multi-resolution convolutional neural network. In: International Journal of Computer Assisted Radiology and Surgery, vol. 16, pp.1077–1087
Foulsham T, Chapman C, Nasiopoulos E, Kingstone A (2014) Top-down and bottom-up aspects of active search in a real-world environment. In: Canadian Journal of Experimental Psychology, vol. 68, pp. 8–19
Li S, Gong Q, Li H, Chen S, Liu Y, Ruan G, Zhu L, Liu L, Chen H (2022) Automatic location scheme of anatomical landmarks in 3d head MRI based on the scale attention hourglass network. Comput Methods Progr Biomed 214:106564
Lester H, Arridge SR (1999) A survey of hierarchical non-linear medical image registration. Pattern Recogn 32(1):129–149
Lefaucheur J-P (2019) Transcranial magnetic stimulation. Handb Clin Neurol 160:559–580
Balconi M (2013) Dorsolateral prefrontal cortex, working memory and episodic memory processes: insight through transcranial magnetic stimulation techniques. Neurosci Bull 29:381–389
Hamid P, Malik BH, Hussain ML (2019) Noninvasive transcranial magnetic stimulation (TMS) in chronic refractory pain: a systematic review. Cureus 11(10):e6019
Sparing R, Buelte D, Meister IG, Pauš T, Fink GR (2008) Transcranial magnetic stimulation and the challenge of coil placement: a comparison of conventional and stereotaxic neuronavigational strategies. Hum Brain Mapp 29(1):82–96
Siebner HR, Hartwigsen G, Kassuba T, Rothwell JC (2009) How does transcranial magnetic stimulation modify neuronal activity in the brain? Implications for studies of cognition. Cortex 45(9):1035–1042
Middlebrooks E, Domingo R, Vivas-Buitrago T, Okromelidze L, Tsuboi T, Wong J, Eisinger R, Almeida L, Burns M, Horn A et al (2020) Neuroimaging advances in deep brain stimulation: review of indications, anatomy, and brain connectomics. Am J Neuroradiol 41(9):1558–1568
Baxter JS, Jannin P (2023) Validation in the age of machine learning: a framework for describing validation with examples in transcranial magnetic stimulation and deep brain stimulation. Intell. -Based Med. 7:100090
Haegelen C, Coupé P, Fonov V, Guizard N, Jannin P, Morandi X, Collins DL (2013) Automated segmentation of basal ganglia and deep brain structures in MRI of Parkinson’s disease. Int J Comput Assist Radiol Surg 8(1):99–110
Long J, Shelhamer E, Darrell T (2015) Fully convolutional networks for semantic segmentation. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 3431–3440
Ronneberger O, Fischer P, Brox T (2015) U-net: convolutional networks for biomedical image segmentation. In: Medical image computing and computer-assisted intervention-MICCAI, 18th International Conference, Munich, Proceedings, Part III 18. Springer 2015, pp. 234–241
Ranftl R, Bochkovskiy A, Koltun V (2021) Vision transformers for dense prediction. In: Proceedings of the IEEE/CVF international conference on computer vision, pp. 12 179–12 188
Hatamizadeh A, Nath V, Tang Y, Yang D, Roth HR, Xu D (2021) Swin unetr: swin transformers for semantic segmentation of brain tumors in MRI images. In: Brainlesion: glioma, multiple sclerosis, stroke and traumatic brain injuries: 7th international workshop, BrainLes. Held in Conjunction with MICCAI 2021, Virtual Event, Revised Selected Papers. Part I. Springer 2022:272–284
Baxter JS, Jannin P (2022) Combining simple interactivity and machine learning: a separable deep learning approach to subthalamic nucleus localization and segmentation in MRI for deep brain stimulation surgical planning. J Med Imaging 9(4):045001
Acknowledgements
We would like to acknowledge SYNEIKA (Rennes, France) for the use of their TMS targeting dataset.
Funding
for EG was received from the Institut des Neurosciences Cliniques de Rennes (INCR) and the Allocations de Recherche Doctorale (ARED) initiative from the Région Bretagne.
Author information
Authors and Affiliations
Contributions
EG and JB designed and carried out the experiments as well as wrote the manuscript. JB and PJ supervised the project.
Corresponding author
Ethics declarations
Conflicts of interest
We have no Conflicts of interest to declare.
Ethical approval
All patient data were collected retrospectively with informed patient consent and approval from the institutional ethics review board.
Consent for publication
We consent to the publication of this article in IJCARS.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Giffard, E., Jannin, P. & Baxter, J. .H. A preliminary exploration into top-down and bottom-up deep-learning approaches to localising neuro-interventional point targets in volumetric MRI. Int J CARS 19, 283–296 (2024). https://doi.org/10.1007/s11548-023-03023-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11548-023-03023-9