1. Cobrawap: a modular cortical wave analysis pipeline for heterogeneous data
- Author
-
Gutzen, R., Bonis, G. D., Grün, S., Davison, A., Paolucci, P. S., Denker, M., Pastorelli, E., Capone, C., Luca, C. D., Mascaro, A. L. A., Resta, F., Pavone, F. S., Sanchez-Vives, M. V., and Mattia, M.
- Abstract
Introduction:An unprecedented richness of data and methodologies enables more detailed access to neural processes but also poses the challenge to combine insights across experiments, species, and measurement techniques. While different experimental recording modalities offer complementary views onto the brain, their data analysis approaches and workflows are often too specific to compare the results rigorously. However, this challenge also promises new avenues of scientific progress. By aligning existing data and analyses from different sources in a reusable workflow we can build a broader basis for meta-studies, contextualization of individual studies, and model validation.Here, we showcase such an analysis pipeline with the application to cortical wave activity in the delta (‘slow waves’) and beta range. Cortical waves can be prominently observed in a variety of heterogeneous data [1,2] and a plethora of analytical methods exist that we aim to interface within a consistent framework: the ‘collaborative brain wave analysis pipeline’ (CobraWap).Methods:The design of CobraWap is based on modular building blocks that provide implementations of analysis methods and processing steps. These blocks are grouped in task-specific stages, e.g., data entry, data processing, trigger detection, wave detection, wave characterization. By letting the pipeline match the input and output format requirements for each of these pipeline components, defining a workflow becomes a matter of selecting a combination of stages and blocks to be applied. This flexibility is employed to converge the heterogeneous data to a common description level of wave activity, from which then common characteristic measures, such as velocity, direction, inter-wave intervals, or wave type classifications, can be derived and quantitatively compared across the data. We demonstrate the versatility of the pipeline with multiple datasets of ECoG [3] and calcium imaging recordings [4] of anesthetized mice, and Utah-array recordings of awake behaving macaques [e.g. 5]. Further, we integrate standard analysis methods from the literature to serve the requirements of a wide range of datasets and research questions. To emphasize the reusability and extendability of each of the pipeline components, the pipeline builds entirely on open-source solutions, such as the workflow manager Snakemake (RRID:SCR_003475), the Neo (RRID:SCR_000634) library for data representation [6], the Elephant (RRID:SCR_003833) analysis toolbox, and the EBRAINS Knowledge Graph (https://kg.ebrains.eu) for capturing outputs of the pipeline execution.Results:The pipeline design promotes the creation of application-tailored and reproducible analysis workflows for many datasets. We demonstrate this “big-data'' approach by investigating dataset-specific parameters across different experiments. For example, we evaluate the influences of the type and dose of anesthesia or the measurement modality and their temporal and spatial resolution on the characteristics of slow waves (e.g., wave velocities) and show that we can replicate corresponding findings from the literature [7,8,9,10].Just as applying the same methods to different data enables a fair comparison between datasets, the pipeline equally enables analyzing the same data with different methods to benchmark their influence on the resulting wave detection and characterization. Finally, we adapt the pipeline for the analysis of beta waves and discuss how the individual elements can be reused, rearranged, or extended to help derive analysis workflows for similar research endeavors and amplify collaborative research.Conclusions:While there are growing efforts in formalizing how neuroscientific data is represented and stored, we here present the benefits of furthermore formalizing the analysis workflows, leveraging the benefits of the diversity in data and methods towards easier collaboration and a cumulative understanding of brain function. REFERENCES[1] Adamantidis, A. R., Herrera C. G., and Gent T. C. (2019) "Oscillating circuitries in the sleeping brain." Nature Reviews Neuroscience 1-17. doi: 10.1038/s41583-019-0223-4[2] Muller, L. et al. (2018). “Cortical Travelling Waves: Mechanisms and Computational Principles.” Nature Reviews Neuroscience 19 (5): 255–68. doi: 10.1038/nrn.2018.20.[3] Sanchez-Vives, M. (2019) “Cortical activity features in transgenic mouse models of cognitive deficits (Williams Beuren Syndrome)” [Data set]. EBRAINS. doi: 10.25493/DZWT-1T8; Sanchez-Vives, M. (2019) "Cortical activity features in transgenic mouse models of cognitive deficits (Williams Beuren Syndrome)" EBRAINS. doi: 10.25493/ANF9-EG3[4] Resta, F., Allegra Mascaro, A. L., & Pavone, F. (2020) "Study of Slow Waves (SWs) propagation through wide-field calcium imaging of the right cortical hemisphere of GCaMP6f mice" EBRAINS. doi: 10.25493/3E6Y-E8G; Resta, F., Allegra Mascaro, A. L., & Pavone, F. (2021) "Study of Slow Waves (SWs) propagation through wide-field calcium imaging of the right cortical hemisphere of GCaMP6f mice (v2)" EBRAINS. doi: 10.25493/QFZK-FXS; Resta, F., [5] Allegra Mascaro, A. L., & Pavone, F. (2020) "Wide-field calcium imaging of the right cortical hemisphere of GCaMP6f mice at different anesthesia levels" EBRAINS. doi: 10.25493/XJR8-QCA[6] Brochier, T. et al. (2018) “Massively Parallel Recordings in Macaque Motor Cortex during an Instructed Delayed Reach-to-Grasp Task.” Scientific Data 5 (1): 180055. doi: 10.1038/sdata.2018.55.[7] Garcia, S. et al. (2014) “Neo: an object model for handling electrophysiology data in multiple formats.” Frontiers in Neuroinformatics 8:10. doi: 10.3389/fninf.2014.00010[8] De Bonis, G. et al. (2019) "Analysis pipeline for extracting features of cortical slow oscillations". Frontiers in Systems Neuroscience 13:70. doi: 10.3389/fnsys.2019.00070[9] Celotto, M. et al. (2020) “Analysis and Model of Cortical Slow Waves Acquired with Optical Techniques”. Methods and Protocols 3.1:14. doi: 10.3390/mps3010014[10] Dasilva, M., et al. (2020). Modulation of cortical slow oscillations and complexity across anesthesia levels. NeuroImage, 224, 117415. doi: 10.1016/j.neuroimage.2020.117415[11] Liang, Y. (2021). “Cortex-Wide Dynamics of Intrinsic Electrical Activities: Propagating Waves and Their Interactions.” Journal of Neuroscience 41 (16): 3665–78. doi: 10.1523/JNEUROSCI.0623-20.2021
- Published
- 2022