التفاصيل البيبلوغرافية
| العنوان: |
Network modeling with gKDR-GMM. |
| المؤلفون: |
Yu Toyoshima, Hirofumi Sato, Daiki Nagata, Manami Kanamori, Moon Sun Jang, Koyo Kuze, Suzu Oe, Takayuki Teramoto, Yuishi Iwasaki, Ryo Yoshida, Takeshi Ishihara, Yuichi Iino |
| سنة النشر: |
2024 |
| المجموعة: |
Smithsonian Institution: Figshare |
| مصطلحات موضوعية: |
Medicine, Cell Biology, Neuroscience, Ecology, Science Policy, Biological Sciences not elsewhere classified, Information Systems not elsewhere classified, variable sensory transmission, strong interactions within, caenorhabditis elegans <, also performed using, scale activity imaging, lapse 3d imaging, core neural circuit, independent component analysis, information flow deduction, extracted common dynamics, successfully decomposed whole, generating realistic whole, div >< p, brain neural network, generate overall dynamics, understanding information flow, component dynamics, brain dynamics, neural activities, 4d imaging, overall relationships, ensemble dynamics, coordinated dynamics |
| الوصف: |
(A) Overview of the gKDR-GMM model. The model learns to predict the target neuron activity (y (t+Δt) ) from the previous activity of presynaptic neurons ( X ) (blue lines represent physical connections). To this end, gKDR reduces the dimension of presynaptic activities X and generates K -dimensional values U representing X , ensuring that sufficient information for predicting y is included in U . GMM models the joint probability of ( U , y ) as a weighted sum of Gaussian distributions fitted to real data. The conditional probability of y is determined from the GMM model and used for prediction. See Methods and S1 Text —File 23 for details. (B) Neural activities obtained from free run simulation by gKDR-GMM. The time span depicted as “real” is actual activity data, identical to Fig 2A whereas the time span depicted as “Repeat 1/2” displays simulation results. The results of two simulation runs are shown following the same order of neurons as Fig 2A . (C) Cross-correlation of neural activities. The left panel shows the cross-correlation of actual activities as in Fig 2B , shown for comparison. The middle and right panels are cross-correlations of the simulation results in (B). Red and blue color show positive and negative correlations, respectively. See S1 Text —File 11A-C for results of all samples. |
| نوع الوثيقة: |
still image |
| اللغة: |
unknown |
| Relation: |
https://figshare.com/articles/figure/Network_modeling_with_gKDR-GMM_/25420639 |
| DOI: |
10.1371/journal.pcbi.1011848.g006 |
| الاتاحة: |
https://doi.org/10.1371/journal.pcbi.1011848.g006 |
| Rights: |
CC BY 4.0 |
| رقم الانضمام: |
edsbas.ED6038DC |
| قاعدة البيانات: |
BASE |