Representing the dynamics of high-dimensional data with non-redundant wavelets

Shanshan Jia; Xingyi Li; Tiejun Huang; Jian K. Liu; Zhaofei Yu

doi:10.1016/j.patter.2021.100424

Representing the dynamics of high-dimensional data with non-redundant wavelets

Shanshan Jia, Xingyi Li, Tiejun Huang, Jian K. Liu^*, Zhaofei Yu^*

^*Corresponding author for this work

Computer Science

Research output: Contribution to journal › Article › peer-review

9 Downloads (Pure)

Abstract

A crucial question in data science is to extract meaningful information embedded in high-dimensional data into a low-dimensional set of features that can represent the original data at different levels. Wavelet analysis is a pervasive method for decomposing time-series signals into a few levels with detailed temporal resolution. However, obtained wavelets are intertwined and over-represented across levels for each sample and across different samples within one population. Here, using neuroscience data of simulated spikes, experimental spikes, calcium imaging signals, and human electrocorticography signals, we leveraged conditional mutual information between wavelets for feature selection. The meaningfulness of selected features was verified to decode stimulus or condition with high accuracy yet using only a small set of features. These results provide a new way of wavelet analysis for extracting essential features of the dynamics of spatiotemporal neural data, which then enables to support novel model design of machine learning with representative features.

Original language	English
Article number	100424
Number of pages	14
Journal	Patterns
Volume	3
Issue number	3
Early online date	6 Jan 2022
DOIs	https://doi.org/10.1016/j.patter.2021.100424
Publication status	Published - 11 Mar 2022

Bibliographical note

Acknowledgments:
We would like to thank Hongbao Jia, Rodrigo Quian Quiroga, and Stefano Panzeri for helpful discussion. This work was supported by the National Natural Science Foundation of China (62176003, 62088102, and 61961130392), the Beijing Major Science and Technology Project (Z191100010618003), the Chongqing Postdoctoral Science Special Foundation (2019LY18), and the Royal Society Newton Advanced Fellowship, UK (NAF-R1-191082).

Keywords

neural coding
wavelet analysis
mutual information
conditional information
feature selection
dimensionality reduction
neural spikes
calcium imaging
ECoG

Access to Document

10.1016/j.patter.2021.100424Licence: Creative Commons: Attribution-NonCommercial-NoDerivs (CC BY-NC-ND)

JiaS2022RepresentingFinal published version, 4.99 MBLicence: Creative Commons: Attribution-NonCommercial-NoDerivs (CC BY-NC-ND)

https://linkinghub.elsevier.com/retrieve/pii/S266638992100307XLicence: Creative Commons: Attribution-NonCommercial-NoDerivs (CC BY-NC-ND)

Cite this

@article{4a51834befec43c8bbcf2b5cba07d823,

title = "Representing the dynamics of high-dimensional data with non-redundant wavelets",

abstract = "A crucial question in data science is to extract meaningful information embedded in high-dimensional data into a low-dimensional set of features that can represent the original data at different levels. Wavelet analysis is a pervasive method for decomposing time-series signals into a few levels with detailed temporal resolution. However, obtained wavelets are intertwined and over-represented across levels for each sample and across different samples within one population. Here, using neuroscience data of simulated spikes, experimental spikes, calcium imaging signals, and human electrocorticography signals, we leveraged conditional mutual information between wavelets for feature selection. The meaningfulness of selected features was verified to decode stimulus or condition with high accuracy yet using only a small set of features. These results provide a new way of wavelet analysis for extracting essential features of the dynamics of spatiotemporal neural data, which then enables to support novel model design of machine learning with representative features.",

keywords = "neural coding, wavelet analysis, mutual information, conditional information, feature selection, dimensionality reduction, neural spikes, calcium imaging, ECoG",

author = "Shanshan Jia and Xingyi Li and Tiejun Huang and Liu, {Jian K.} and Zhaofei Yu",

note = "Acknowledgments: We would like to thank Hongbao Jia, Rodrigo Quian Quiroga, and Stefano Panzeri for helpful discussion. This work was supported by the National Natural Science Foundation of China (62176003, 62088102, and 61961130392), the Beijing Major Science and Technology Project (Z191100010618003), the Chongqing Postdoctoral Science Special Foundation (2019LY18), and the Royal Society Newton Advanced Fellowship, UK (NAF-R1-191082).",

year = "2022",

month = mar,

day = "11",

doi = "10.1016/j.patter.2021.100424",

language = "English",

volume = "3",

journal = "Patterns",

issn = "2666-3899",

publisher = "Cell Press",

number = "3",

}

TY - JOUR

T1 - Representing the dynamics of high-dimensional data with non-redundant wavelets

AU - Jia, Shanshan

AU - Li, Xingyi

AU - Huang, Tiejun

AU - Liu, Jian K.

AU - Yu, Zhaofei

N1 - Acknowledgments: We would like to thank Hongbao Jia, Rodrigo Quian Quiroga, and Stefano Panzeri for helpful discussion. This work was supported by the National Natural Science Foundation of China (62176003, 62088102, and 61961130392), the Beijing Major Science and Technology Project (Z191100010618003), the Chongqing Postdoctoral Science Special Foundation (2019LY18), and the Royal Society Newton Advanced Fellowship, UK (NAF-R1-191082).

PY - 2022/3/11

Y1 - 2022/3/11

N2 - A crucial question in data science is to extract meaningful information embedded in high-dimensional data into a low-dimensional set of features that can represent the original data at different levels. Wavelet analysis is a pervasive method for decomposing time-series signals into a few levels with detailed temporal resolution. However, obtained wavelets are intertwined and over-represented across levels for each sample and across different samples within one population. Here, using neuroscience data of simulated spikes, experimental spikes, calcium imaging signals, and human electrocorticography signals, we leveraged conditional mutual information between wavelets for feature selection. The meaningfulness of selected features was verified to decode stimulus or condition with high accuracy yet using only a small set of features. These results provide a new way of wavelet analysis for extracting essential features of the dynamics of spatiotemporal neural data, which then enables to support novel model design of machine learning with representative features.

AB - A crucial question in data science is to extract meaningful information embedded in high-dimensional data into a low-dimensional set of features that can represent the original data at different levels. Wavelet analysis is a pervasive method for decomposing time-series signals into a few levels with detailed temporal resolution. However, obtained wavelets are intertwined and over-represented across levels for each sample and across different samples within one population. Here, using neuroscience data of simulated spikes, experimental spikes, calcium imaging signals, and human electrocorticography signals, we leveraged conditional mutual information between wavelets for feature selection. The meaningfulness of selected features was verified to decode stimulus or condition with high accuracy yet using only a small set of features. These results provide a new way of wavelet analysis for extracting essential features of the dynamics of spatiotemporal neural data, which then enables to support novel model design of machine learning with representative features.

KW - neural coding

KW - wavelet analysis

KW - mutual information

KW - conditional information

KW - feature selection

KW - dimensionality reduction

KW - neural spikes

KW - calcium imaging

KW - ECoG

U2 - 10.1016/j.patter.2021.100424

DO - 10.1016/j.patter.2021.100424

M3 - Article

SN - 2666-3899

VL - 3

JO - Patterns

JF - Patterns

IS - 3

M1 - 100424

ER -

Representing the dynamics of high-dimensional data with non-redundant wavelets

Abstract

Bibliographical note

Keywords

Access to Document

Fingerprint

Cite this