scholarly journals Deep-learning-based whole-brain imaging at single-neuron resolution

2020 ◽  
Vol 11 (7) ◽  
pp. 3567
Author(s):  
Kefu Ning ◽  
Xiaoyu Zhang ◽  
Xuefei Gao ◽  
Tao Jiang ◽  
He Wang ◽  
...  
Author(s):  
Kefu Ning ◽  
Xiaoyu Zhang ◽  
Xuefei Gao ◽  
Tao Jiang ◽  
He Wang ◽  
...  

2018 ◽  
Author(s):  
Chentao Wen ◽  
Takuya Miura ◽  
Yukako Fujie ◽  
Takayuki Teramoto ◽  
Takeshi Ishihara ◽  
...  

AbstractThe brain is a complex system that operates based on coordinated neuronal activities. Brain-wide cellular calcium imaging techniques have quickly advanced in recent years and become powerful tools for understanding the neuronal activities of small animal models. The whole brain imaging generally requires to extract the neuronal activities from three-dimensional (3D) image series. Unfortunately, the 3D image series are obtained under imaging conditions different among laboratories and extracting neuronal activities from the data requires multiple processes. Therefore researchers need to develop their own software, which has prevented the application of whole-brain imaging experiments in more laboratories. Here, we combined traditional image processing techniques with the powerful deep-learning method which can be flexibly modified to fit 3D image data in the nematode Caenorhabditis elegans obtained under different conditions. We first trained the 3D U-net deep network to classify each pixel into cell and non-cell categories. Cells merged as a whole region were further separated into individual cells by watershed segmentation. The cells were then tracked in 3D space over time with the combination of a feedforward network and a point set registration method to use local and global relative positions of the cells, respectively. Remarkably, one manually annotated 3D image combined with data augmentation was sufficient for training the deep networks to obtain satisfactory tracking results. Our method correctly tracked more than 98% of neurons in three different image datasets and successfully extracted brain-wide neuronal activities. Our method worked well even when the sampling rate was reduced: 86% correct in case 4/5 frames were removed, and when artificial noise was added into the raw images: 91% correct in case 35 times of background-level noise was added. Our results proved that deep learning is widely applicable to different datasets and can help us in establishing a flexible pipeline for extracting whole brain activities.


Cell Reports ◽  
2021 ◽  
Vol 34 (5) ◽  
pp. 108709
Author(s):  
Xiaojun Wang ◽  
Hanqing Xiong ◽  
Yurong Liu ◽  
Tao Yang ◽  
Anan Li ◽  
...  

2018 ◽  
Vol 527 (13) ◽  
pp. 2122-2145 ◽  
Author(s):  
Jennifer D. Whitesell ◽  
Alex R. Buckley ◽  
Joseph E. Knox ◽  
Leonard Kuan ◽  
Nile Graddis ◽  
...  

2012 ◽  
Vol 107 (10) ◽  
pp. 2853-2865 ◽  
Author(s):  
Ji-Wei He ◽  
Fenghua Tian ◽  
Hanli Liu ◽  
Yuan Bo Peng

While near-infrared (NIR) spectroscopy has been increasingly used to detect stimulated brain activities with an advantage of dissociating regional oxy- and deoxyhemoglobin concentrations simultaneously, it has not been utilized much in pain research. Here, we investigated and demonstrated the feasibility of using this technique to obtain whole brain hemodynamics in rats and speculated on the functional relevance of the NIR-based hemodynamic signals during pain processing. NIR signals were emitted and collected using a 26-optodes array on rat's dorsal skull surface after the removal of skin. Following the subcutaneous injection of formalin (50 μl, 3%) into a hindpaw, several isolable brain regions showed hemodynamic changes, including the anterior cingulate cortex, primary/secondary somatosensory cortexes, thalamus, and periaqueductal gray ( n = 6). Time courses of hemodynamic changes in respective regions matched with the well-documented biphasic excitatory response. Surprisingly, an atypical pattern (i.e., a decrease in oxyhemoglobin concentration with a concomitant increase in deoxyhemoglobin concentration) was seen in phase II. In a separate group of rats with innocuous brush and noxious pinch of the same area ( n = 11), results confirmed that the atypical pattern occurred more likely in the presence of nociception than nonpainful stimulation, suggesting it as a physiological substrate when the brain processes pain. In conclusion, the NIR whole brain imaging provides a useful alternative to study pain in vivo using small-animal models. Our results support the notion that neurovascular response patterns depend on stimuli, bringing attention to the interpretation of vascular-based neuroimaging data in studies of pain.


Author(s):  
Matt Carter ◽  
Jennifer C. Shieh
Keyword(s):  

Author(s):  
Rayyan Manwar ◽  
Xin Li ◽  
Sadreddin Mahmoodkalayeh ◽  
Eishi Asano ◽  
Dongxiao Zhu ◽  
...  
Keyword(s):  

2019 ◽  
Vol 2019 ◽  
pp. 1-9
Author(s):  
Jinlong Hu ◽  
Yuezhen Kuang ◽  
Bin Liao ◽  
Lijie Cao ◽  
Shoubin Dong ◽  
...  

Deep learning models have been successfully applied to the analysis of various functional MRI data. Convolutional neural networks (CNN), a class of deep neural networks, have been found to excel at extracting local meaningful features based on their shared-weights architecture and space invariance characteristics. In this study, we propose M2D CNN, a novel multichannel 2D CNN model, to classify 3D fMRI data. The model uses sliced 2D fMRI data as input and integrates multichannel information learned from 2D CNN networks. We experimentally compared the proposed M2D CNN against several widely used models including SVM, 1D CNN, 2D CNN, 3D CNN, and 3D separable CNN with respect to their performance in classifying task-based fMRI data. We tested M2D CNN against six models as benchmarks to classify a large number of time-series whole-brain imaging data based on a motor task in the Human Connectome Project (HCP). The results of our experiments demonstrate the following: (i) convolution operations in the CNN models are advantageous for high-dimensional whole-brain imaging data classification, as all CNN models outperform SVM; (ii) 3D CNN models achieve higher accuracy than 2D CNN and 1D CNN model, but 3D CNN models are computationally costly as any extra dimension is added in the input; (iii) the M2D CNN model proposed in this study achieves the highest accuracy and alleviates data overfitting given its smaller number of parameters as compared with 3D CNN.


Sign in / Sign up

Export Citation Format

Share Document