Browsing by Subject "Neural Networks, Computer"

Filter results by typing the first few letters
Now showing 1 - 11 of 11
  • Thumbnail Image
    ItemOpen Access
    A Deep Learning Approach for Rapid and Generalizable Denoising of Photon-Counting Micro-CT Images.
    (Tomography (Ann Arbor, Mich.), 2023-07) Nadkarni, Rohan; Clark, Darin P; Allphin, Alex J; Badea, Cristian T
    Photon-counting CT (PCCT) is powerful for spectral imaging and material decomposition but produces noisy weighted filtered backprojection (wFBP) reconstructions. Although iterative reconstruction effectively denoises these images, it requires extensive computation time. To overcome this limitation, we propose a deep learning (DL) model, UnetU, which quickly estimates iterative reconstruction from wFBP. Utilizing a 2D U-net convolutional neural network (CNN) with a custom loss function and transformation of wFBP, UnetU promotes accurate material decomposition across various photon-counting detector (PCD) energy threshold settings. UnetU outperformed multi-energy non-local means (ME NLM) and a conventional denoising CNN called UnetwFBP in terms of root mean square error (RMSE) in test set reconstructions and their respective matrix inversion material decompositions. Qualitative results in reconstruction and material decomposition domains revealed that UnetU is the best approximation of iterative reconstruction. In reconstructions with varying undersampling factors from a high dose ex vivo scan, UnetU consistently gave higher structural similarity (SSIM) and peak signal-to-noise ratio (PSNR) to the fully sampled iterative reconstruction than ME NLM and UnetwFBP. This research demonstrates UnetU's potential as a fast (i.e., 15 times faster than iterative reconstruction) and generalizable approach for PCCT denoising, holding promise for advancing preclinical PCCT research.
  • Thumbnail Image
    ItemOpen Access
    A new open-access platform for measuring and sharing mTBI data.
    (Scientific reports, 2021-04) Domel, August G; Raymond, Samuel J; Giordano, Chiara; Liu, Yuzhe; Yousefsani, Seyed Abdolmajid; Fanton, Michael; Cecchi, Nicholas J; Vovk, Olga; Pirozzi, Ileana; Kight, Ali; Avery, Brett; Boumis, Athanasia; Fetters, Tyler; Jandu, Simran; Mehring, William M; Monga, Sam; Mouchawar, Nicole; Rangel, India; Rice, Eli; Roy, Pritha; Sami, Sohrab; Singh, Heer; Wu, Lyndia; Kuo, Calvin; Zeineh, Michael; Grant, Gerald; Camarillo, David B
    Despite numerous research efforts, the precise mechanisms of concussion have yet to be fully uncovered. Clinical studies on high-risk populations, such as contact sports athletes, have become more common and give insight on the link between impact severity and brain injury risk through the use of wearable sensors and neurological testing. However, as the number of institutions operating these studies grows, there is a growing need for a platform to share these data to facilitate our understanding of concussion mechanisms and aid in the development of suitable diagnostic tools. To that end, this paper puts forth two contributions: (1) a centralized, open-access platform for storing and sharing head impact data, in collaboration with the Federal Interagency Traumatic Brain Injury Research informatics system (FITBIR), and (2) a deep learning impact detection algorithm (MiGNet) to differentiate between true head impacts and false positives for the previously biomechanically validated instrumented mouthguard sensor (MiG2.0), all of which easily interfaces with FITBIR. We report 96% accuracy using MiGNet, based on a neural network model, improving on previous work based on Support Vector Machines achieving 91% accuracy, on an out of sample dataset of high school and collegiate football head impacts. The integrated MiG2.0 and FITBIR system serve as a collaborative research tool to be disseminated across multiple institutions towards creating a standardized dataset for furthering the knowledge of concussion biomechanics.
  • Thumbnail Image
    ItemOpen Access
    Attractor Dynamics in Networks with Learning Rules Inferred from In Vivo Data.
    (Neuron, 2018-07) Pereira, Ulises; Brunel, Nicolas
    The attractor neural network scenario is a popular scenario for memory storage in the association cortex, but there is still a large gap between models based on this scenario and experimental data. We study a recurrent network model in which both learning rules and distribution of stored patterns are inferred from distributions of visual responses for novel and familiar images in the inferior temporal cortex (ITC). Unlike classical attractor neural network models, our model exhibits graded activity in retrieval states, with distributions of firing rates that are close to lognormal. Inferred learning rules are close to maximizing the number of stored patterns within a family of unsupervised Hebbian learning rules, suggesting that learning rules in ITC are optimized to store a large number of attractor states. Finally, we show that there exist two types of retrieval states: one in which firing rates are constant in time and another in which firing rates fluctuate chaotically.
  • Thumbnail Image
    ItemOpen Access
    Automated multimodal segmentation of acute ischemic stroke lesions on clinical MR images.
    (Magnetic resonance imaging, 2022-10) Moon, Hae Sol; Heffron, Lindsay; Mahzarnia, Ali; Obeng-Gyasi, Barnabas; Holbrook, Matthew; Badea, Cristian T; Feng, Wuwei; Badea, Alexandra
    Magnetic resonance (MR) imaging (MRI) is commonly used to diagnose, assess and monitor stroke. Accurate and timely segmentation of stroke lesions provides the anatomico-structural information that can aid physicians in predicting prognosis, as well as in decision making and triaging for various rehabilitation strategies. To segment stroke lesions, MR protocols, including diffusion-weighted imaging (DWI) and T2-weighted fluid attenuated inversion recovery (FLAIR) are often utilized. These imaging sequences are usually acquired with different spatial resolutions due to time constraints. Within the same image, voxels may be anisotropic, with reduced resolution along slice direction for diffusion scans in particular. In this study, we evaluate the ability of 2D and 3D U-Net Convolutional Neural Network (CNN) architectures to segment ischemic stroke lesions using single contrast (DWI) and dual contrast images (T2w FLAIR and DWI). The predicted segmentations correlate with post-stroke motor outcome measured by the National Institutes of Health Stroke Scale (NIHSS) and Fugl-Meyer Upper Extremity (FM-UE) index based on the lesion loads overlapping the corticospinal tracts (CST), which is a neural substrate for motor movement and function. Although the four methods performed similarly, the 2D multimodal U-Net achieved the best results with a mean Dice of 0.737 (95% CI: 0.705, 0.769) and a relatively high correlation between the weighted lesion load and the NIHSS scores (both at baseline and at 90 days). A monotonically constrained quintic polynomial regression yielded R2 = 0.784 and 0.875 for weighted lesion load versus baseline and 90-Days NIHSS respectively, and better corrected Akaike information criterion (AICc) scores than those of the linear regression. In addition, using the quintic polynomial regression model to regress the weighted lesion load to the 90-Days FM-UE score results in an R2 of 0.570 with a better AICc score than that of the linear regression. Our results suggest that the multi-contrast information enhanced the accuracy of the segmentation and the prediction accuracy for upper extremity motor outcomes. Expanding the training dataset to include different types of stroke lesions and more data points will help add a temporal longitudinal aspect and increase the accuracy. Furthermore, adding patient-specific data may improve the inference about the relationship between imaging metrics and functional outcomes.
  • Thumbnail Image
    ItemOpen Access
    Cerebellar learning using perturbations.
    (eLife, 2018-11-12) Bouvier, Guy; Aljadeff, Johnatan; Clopath, Claudia; Bimbard, Célian; Ranft, Jonas; Blot, Antonin; Nadal, Jean-Pierre; Brunel, Nicolas; Hakim, Vincent; Barbour, Boris
    The cerebellum aids the learning of fast, coordinated movements. According to current consensus, erroneously active parallel fibre synapses are depressed by complex spikes signalling movement errors. However, this theory cannot solve the credit assignment problem of processing a global movement evaluation into multiple cell-specific error signals. We identify a possible implementation of an algorithm solving this problem, whereby spontaneous complex spikes perturb ongoing movements, create eligibility traces and signal error changes guiding plasticity. Error changes are extracted by adaptively cancelling the average error. This framework, stochastic gradient descent with estimated global errors (SGDEGE), predicts synaptic plasticity rules that apparently contradict the current consensus but were supported by plasticity experiments in slices from mice under conditions designed to be physiological, highlighting the sensitivity of plasticity studies to experimental conditions. We analyse the algorithm's convergence and capacity. Finally, we suggest SGDEGE may also operate in the basal ganglia.
  • Thumbnail Image
    ItemOpen Access
    Characteristics of sequential activity in networks with temporally asymmetric Hebbian learning.
    (Proceedings of the National Academy of Sciences of the United States of America, 2020-11-11) Gillett, Maxwell; Pereira, Ulises; Brunel, Nicolas
    Sequential activity has been observed in multiple neuronal circuits across species, neural structures, and behaviors. It has been hypothesized that sequences could arise from learning processes. However, it is still unclear whether biologically plausible synaptic plasticity rules can organize neuronal activity to form sequences whose statistics match experimental observations. Here, we investigate temporally asymmetric Hebbian rules in sparsely connected recurrent rate networks and develop a theory of the transient sequential activity observed after learning. These rules transform a sequence of random input patterns into synaptic weight updates. After learning, recalled sequential activity is reflected in the transient correlation of network activity with each of the stored input patterns. Using mean-field theory, we derive a low-dimensional description of the network dynamics and compute the storage capacity of these networks. Multiple temporal characteristics of the recalled sequential activity are consistent with experimental observations. We find that the degree of sparseness of the recalled sequences can be controlled by nonlinearities in the learning rule. Furthermore, sequences maintain robust decoding, but display highly labile dynamics, when synaptic connectivity is continuously modified due to noise or storage of other patterns, similar to recent observations in hippocampus and parietal cortex. Finally, we demonstrate that our results also hold in recurrent networks of spiking neurons with separate excitatory and inhibitory populations.
  • Thumbnail Image
    ItemOpen Access
    Is cortical connectivity optimized for storing information?
    (Nature neuroscience, 2016-05) Brunel, Nicolas
    Cortical networks are thought to be shaped by experience-dependent synaptic plasticity. Theoretical studies have shown that synaptic plasticity allows a network to store a memory of patterns of activity such that they become attractors of the dynamics of the network. Here we study the properties of the excitatory synaptic connectivity in a network that maximizes the number of stored patterns of activity in a robust fashion. We show that the resulting synaptic connectivity matrix has the following properties: it is sparse, with a large fraction of zero synaptic weights ('potential' synapses); bidirectionally coupled pairs of neurons are over-represented in comparison to a random network; and bidirectionally connected pairs have stronger synapses on average than unidirectionally connected pairs. All these features reproduce quantitatively available data on connectivity in cortex. This suggests synaptic connectivity in cortex is optimized to store a large number of attractor states in a robust fashion.
  • Thumbnail Image
    ItemOpen Access
    Multi-label annotation of text reports from computed tomography of the chest, abdomen, and pelvis using deep learning.
    (BMC medical informatics and decision making, 2022-04) D'Anniballe, Vincent M; Tushar, Fakrul Islam; Faryna, Khrystyna; Han, Songyue; Mazurowski, Maciej A; Rubin, Geoffrey D; Lo, Joseph Y

    Background

    There is progress to be made in building artificially intelligent systems to detect abnormalities that are not only accurate but can handle the true breadth of findings that radiologists encounter in body (chest, abdomen, and pelvis) computed tomography (CT). Currently, the major bottleneck for developing multi-disease classifiers is a lack of manually annotated data. The purpose of this work was to develop high throughput multi-label annotators for body CT reports that can be applied across a variety of abnormalities, organs, and disease states thereby mitigating the need for human annotation.

    Methods

    We used a dictionary approach to develop rule-based algorithms (RBA) for extraction of disease labels from radiology text reports. We targeted three organ systems (lungs/pleura, liver/gallbladder, kidneys/ureters) with four diseases per system based on their prevalence in our dataset. To expand the algorithms beyond pre-defined keywords, attention-guided recurrent neural networks (RNN) were trained using the RBA-extracted labels to classify reports as being positive for one or more diseases or normal for each organ system. Alternative effects on disease classification performance were evaluated using random initialization or pre-trained embedding as well as different sizes of training datasets. The RBA was tested on a subset of 2158 manually labeled reports and performance was reported as accuracy and F-score. The RNN was tested against a test set of 48,758 reports labeled by RBA and performance was reported as area under the receiver operating characteristic curve (AUC), with 95% CIs calculated using the DeLong method.

    Results

    Manual validation of the RBA confirmed 91-99% accuracy across the 15 different labels. Our models extracted disease labels from 261,229 radiology reports of 112,501 unique subjects. Pre-trained models outperformed random initialization across all diseases. As the training dataset size was reduced, performance was robust except for a few diseases with a relatively small number of cases. Pre-trained classification AUCs reached > 0.95 for all four disease outcomes and normality across all three organ systems.

    Conclusions

    Our label-extracting pipeline was able to encompass a variety of cases and diseases in body CT reports by generalizing beyond strict rules with exceptional accuracy. The method described can be easily adapted to enable automated labeling of hospital-scale medical data sets for training image-based disease classifiers.
  • Thumbnail Image
    ItemOpen Access
    Robust deep learning classification of adamantinomatous craniopharyngioma from limited preoperative radiographic images.
    (Scientific reports, 2020-10) Prince, Eric W; Whelan, Ros; Mirsky, David M; Stence, Nicholas; Staulcup, Susan; Klimo, Paul; Anderson, Richard CE; Niazi, Toba N; Grant, Gerald; Souweidane, Mark; Johnston, James M; Jackson, Eric M; Limbrick, David D; Smith, Amy; Drapeau, Annie; Chern, Joshua J; Kilburn, Lindsay; Ginn, Kevin; Naftel, Robert; Dudley, Roy; Tyler-Kabara, Elizabeth; Jallo, George; Handler, Michael H; Jones, Kenneth; Donson, Andrew M; Foreman, Nicholas K; Hankinson, Todd C
    Deep learning (DL) is a widely applied mathematical modeling technique. Classically, DL models utilize large volumes of training data, which are not available in many healthcare contexts. For patients with brain tumors, non-invasive diagnosis would represent a substantial clinical advance, potentially sparing patients from the risks associated with surgical intervention on the brain. Such an approach will depend upon highly accurate models built using the limited datasets that are available. Herein, we present a novel genetic algorithm (GA) that identifies optimal architecture parameters using feature embeddings from state-of-the-art image classification networks to identify the pediatric brain tumor, adamantinomatous craniopharyngioma (ACP). We optimized classification models for preoperative Computed Tomography (CT), Magnetic Resonance Imaging (MRI), and combined CT and MRI datasets with demonstrated test accuracies of 85.3%, 83.3%, and 87.8%, respectively. Notably, our GA improved baseline model performance by up to 38%. This work advances DL and its applications within healthcare by identifying optimized networks in small-scale data contexts. The proposed system is easily implementable and scalable for non-invasive computer-aided diagnosis, even for uncommon diseases.
  • Thumbnail Image
    ItemOpen Access
    Storing structured sparse memories in a multi-modular cortical network model.
    (Journal of computational neuroscience, 2016-04) Dubreuil, Alexis M; Brunel, Nicolas
    We study the memory performance of a class of modular attractor neural networks, where modules are potentially fully-connected networks connected to each other via diluted long-range connections. On this anatomical architecture we store memory patterns of activity using a Willshaw-type learning rule. P patterns are split in categories, such that patterns of the same category activate the same set of modules. We first compute the maximal storage capacity of these networks. We then investigate their error-correction properties through an exhaustive exploration of parameter space, and identify regions where the networks behave as an associative memory device. The crucial parameters that control the retrieval abilities of the network are (1) the ratio between the number of synaptic contacts of long- and short-range origins (2) the number of categories in which a module is activated and (3) the amount of local inhibition. We discuss the relationship between our model and networks of cortical patches that have been observed in different cortical areas.
  • Thumbnail Image
    ItemOpen Access
    Understanding the relationships between spike rate and delta/gamma frequency bands of LFPs and EEGs using a local cortical network model.
    (NeuroImage, 2010-09) Mazzoni, Alberto; Whittingstall, Kevin; Brunel, Nicolas; Logothetis, Nikos K; Panzeri, Stefano
    Despite the widespread use of EEGs to measure the large-scale dynamics of the human brain, little is known on how the dynamics of EEGs relates to that of the underlying spike rates of cortical neurons. However, progress was made by recent neurophysiological experiments reporting that EEG delta-band phase and gamma-band amplitude reliably predict some complementary aspects of the time course of spikes of visual cortical neurons. To elucidate the mechanisms behind these findings, here we hypothesize that the EEG delta phase reflects shifts of local cortical excitability arising from slow fluctuations in the network input due to entrainment to sensory stimuli or to fluctuations in ongoing activity, and that the resulting local excitability fluctuations modulate both the spike rate and the engagement of excitatory-inhibitory loops producing gamma-band oscillations. We quantitatively tested these hypotheses by simulating a recurrent network of excitatory and inhibitory neurons stimulated with dynamic inputs presenting temporal regularities similar to that of thalamic responses during naturalistic visual stimulation and during spontaneous activity. The network model reproduced in detail the experimental relationships between spike rate and EEGs, and suggested that the complementariness of the prediction of spike rates obtained from EEG delta phase or gamma amplitude arises from nonlinearities in the engagement of excitatory-inhibitory loops and from temporal modulations in the amplitude of the network input, which respectively limit the predictability of spike rates from gamma amplitude or delta phase alone. The model suggested also ways to improve and extend current algorithms for online prediction of spike rates from EEGs.