Adaptabe Workflows for Neural Activity Analysis

Adaptable workflows
for neural activity analysis
in an open-source environment

Robin Gutzen

2023-06-21 | International Forum on Neural Engineering & Brain Technologies

r.gutzen@fz-juelich.de

@rgutzen@mstdn.social

Cortical Wave Activty

propagation of activity (phase) in spatially coordinated patterns

data from Chen et al. (2002) Sci. Data

Cortical Wave Activty

(in different frequency ranges)

is thought to ...

degree of consciousness

Koch et al. (2016)

Dasilva et al. (2021)

Pazienti et al. (2022)

indicate

attention

Ermentrout & Kleinfeld (2001)

Fries et al. (2001)

Lakatos et al. (2008)

Jensen et al. (2021)

Bhattacharya et al. (2022)

gate

information processing

Fries (2005)

Roland et al. (2006)

Xu et al. (2007)

van Kerkoerle et al. (2014)

Michalareas et al. (2016)

coordinate

task performance

Zhang et al. (2018)

Davis et al. (2020)

predict

working memory

Jutras et al. (2013)

Bhattacharya et al. (2022)

maintain

neurological disorders

Alexander et al. (2009)

Ferrarelli et al. (2007)

Dasilva et al, (2020)

Sato et al. (2022)

potentially mark

classification learning

Tonielli et al. (2022)

improve

sleep quality

Massimini et al. (2004)

underlie

Huber et al. (2004)

Muller et al. (2016)

Botella-Soler et al. (2012)

Massimini et al. (2004)

Challenge | Lack of Comparability

measuring slow waves (<1Hz)

during sleep in humans

using the equivalent methods

(wave definition, filters, detection thresholds)

using EEG recordings

velocity = 2.7 +- 0.2 m/s

velocity = 1.0 +- 0.2 m/s

effect of

implementation or biology

Challenge | Heterogeneous Data Sources & Non-reusable Software

Data providing viewpoints from various

measurement techniques
spatial and temporal scales
species, and brain areas

Analysis tools should be

reusable
extendable & combinable
maintainable

inset figure references in Gutzen et al . (2022)

... and to leverage this data richness

Challenge | Puzzle Pieces vs. Building Blocks

model for representing electrophysiology data intefacing to various file formats

community-centered library for the analysis of electrophysiological data

workflow management system is a tool to create reproducible and scalable data analyses

open research infrastructure that gathers data, tools and computing facilities

Modularity

elements are combinable in multiple ways

Adaptability

elements can be added, removed, or changed

Reproducibility

elements are individually maintainable

Reusability

elements and parts are useful on their own

Versatility

usability can be expanded beyond the inital scope

Approach:

Creating modular shared components
operating on common standardized descriptions.

Goal:

Providing reusable analysis workflows for
relating knowledge across neuroscience domains.

common
brain wave characterizations

Approach | Common Description Levels

EEG

M EG

ECoG

Calcium Imaging

Implanted

Arrays

Model

Simulation

...

spatio-temporal

activity data

cobrawap.readthedocs.io

Approach | Pipeline of Modular Components

EEG

M EG

ECoG

Calcium

Imaging

Implanted

Arrays

Model

Simulation

...

ECoG

Calcium

Imaging

ECoG

Calcium

Imaging

Application | Meta-Study across Data Sources

Gutzen et al. (2022) arxiv

using 60 open-access datasets of anesthetized mice

Model

Simulation

Calcium

Imaging

Application | Calibration & Validation of Model Simulations

Capone et al. (2023) Comm. Bio

Application | Benchmarking of Analysis Methods

Gutzen et al. (2022) arxiv

hilbert phase

minima

different methods - same data

Implanted

Electrode

Arrays

Application | Linking Cortical Waves to Behavior

adapting to different measurement technique, species, wave types

Gutzen et al. (2023) in progress

Implanted

Electode

Arrays

Application | Linking Cortical Waves to Spiking Activity

Figure adapted from Sven Krausse

extending previous analysis workflows with additional activity modalities

Calcium

Imaging

Application | Measuring Effects of Stroke on Sleep

porting workflows to applications in external labs

Figure adapted from Ben Miao & Eric Landsness

Summary

Comparability across Data Sources

Calibration & Validation

Method Benchmarking

Waves <-> Behavior

Waves <-> Spike Patterns

Waves <-> Disease States

Adaption to

Different Wave Types

EEG

M EG

ECoG

Calcium

Imaging

Implanted

Arrays

Model

Simulation

...

Conclusion

It is relevant to evaluate if specific neuroscience challenges are due to constraints in biology, technology, or philosophy and address them accordingly.

slides at rgutzen.github.io/publications/slides/2023-06-21_NBT

Robin Gutzen

(r.gutzen@fz-juelich.de)

There are a lot of new insights possible, just by developing the appropriate computational tools. For research software, in particular, closer collaborative developments of academia and industry would be mutually beneficial.

"Newton said: 'If I have seen further than others, it is because I've stood on the shoulders of giants.'' These days we stand on each other's feet! You should do your job in such a fashion that others can build on top of it. [..] The essence of science is cumulative."

- Richard Hamming (1986)

Thanks to all Contributers and Collaborators