Longitudinal tracking of neuronal activity from the same cells in the developing brain using Track2p
Curation statements for this article:-
Curated by eLife
eLife Assessment
This important study presents a new method for longitudinally tracking cells in two-photon imaging data that addresses the specific challenges of imaging neurons in the developing cortex. It provides compelling evidence demonstrating reliable longitudinal identification of neurons across the second postnatal week in mice. The study should be of interest to development neuroscientists engaged in population-level recordings using two-photon imaging.
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (eLife)
Abstract
Abstract
Understanding cortical circuit development requires tracking neuronal activity across days in the growing brain. While in vivo calcium imaging now enables such longitudinal studies, automated tools for reliably tracking large populations of neurons across sessions remain limited. Here, we present a novel cell-tracking method based on sequential image registration, validated on calcium imaging data from the barrel cortex of mouse pups over one postnatal week. Our approach enables robust long-term analysis of several hundreds of individual neurons, allowing quantification of neuronal dynamics and representational stability over time. Using this method, we identified a key developmental transition in neuronal activity statistics, marking the emergence of arousal state modulation. Beyond this key finding, our method provides an essential tool for tracking developmental trajectories of individual neurons, which could help identify potential deviations associated with neurodevelopmental disorders.
Article activity feed
-
eLife Assessment
This important study presents a new method for longitudinally tracking cells in two-photon imaging data that addresses the specific challenges of imaging neurons in the developing cortex. It provides compelling evidence demonstrating reliable longitudinal identification of neurons across the second postnatal week in mice. The study should be of interest to development neuroscientists engaged in population-level recordings using two-photon imaging.
-
Reviewer #1 (Public review):
Summary:
This manuscript presents a compelling and innovative approach that combines Track2p neuronal tracking with advanced analytical methods to investigate early postnatal brain development. The work provides a powerful framework for exploring complex developmental processes such as the emergence of sensory representations, cognitive functions, and activity-dependent circuit formation. By enabling the tracking of the same neurons over extended developmental periods, this methodology sets the stage for mechanistic insights that were previously inaccessible.
Strengths:
(1) Innovative Methodology:
The integration of Track2p with longitudinal calcium imaging offers a unique capability to follow individual neurons across critical developmental windows.(2) High Conceptual Impact:
The manuscript outlines a …Reviewer #1 (Public review):
Summary:
This manuscript presents a compelling and innovative approach that combines Track2p neuronal tracking with advanced analytical methods to investigate early postnatal brain development. The work provides a powerful framework for exploring complex developmental processes such as the emergence of sensory representations, cognitive functions, and activity-dependent circuit formation. By enabling the tracking of the same neurons over extended developmental periods, this methodology sets the stage for mechanistic insights that were previously inaccessible.
Strengths:
(1) Innovative Methodology:
The integration of Track2p with longitudinal calcium imaging offers a unique capability to follow individual neurons across critical developmental windows.(2) High Conceptual Impact:
The manuscript outlines a clear path for using this approach to study foundational developmental questions, such as how early neuronal activity shapes later functional properties and network assembly.(3) Future Experimental Potential:
The authors convincingly argue for the feasibility of extending this tracking into adulthood and combining it with targeted manipulations, which could significantly advance our understanding of causality in developmental processes.(4) Broad Applicability:
The proposed framework can be adapted to a wide range of experimental designs and questions, making it a valuable resource for the field.Weaknesses:
No major weaknesses were identified by this reviewer. The manuscript is conceptually strong and methodologically sound. Future studies will need to address potential technical limitations of long-term tracking, but this does not detract from the current work's significance and clarity of vision.
-
Reviewer #2 (Public review):
Summary:
The manuscript by Majnik and colleagues introduces "Track2p", a new tool designed to track neurons across imaging sessions of two-photon calcium imaging in developing mice. The method addresses the challenge of tracking cells in the growing brain of developing mice. The authors showed that "Track2p" successfully tracks hundreds of neurons in the barrel cortex across multiple days during the second postnatal week. This enabled the identification of the emergence of behavioral state modulation and desynchronization of spontaneous network activity around postnatal day 11.
Strengths:
The manuscript is well written, and the analysis pipeline is clearly described. Moreover, the dataset used for validation is of high quality, considering the technical challenges associated with longitudinal two-photon …
Reviewer #2 (Public review):
Summary:
The manuscript by Majnik and colleagues introduces "Track2p", a new tool designed to track neurons across imaging sessions of two-photon calcium imaging in developing mice. The method addresses the challenge of tracking cells in the growing brain of developing mice. The authors showed that "Track2p" successfully tracks hundreds of neurons in the barrel cortex across multiple days during the second postnatal week. This enabled the identification of the emergence of behavioral state modulation and desynchronization of spontaneous network activity around postnatal day 11.
Strengths:
The manuscript is well written, and the analysis pipeline is clearly described. Moreover, the dataset used for validation is of high quality, considering the technical challenges associated with longitudinal two-photon recordings in mouse pups. The authors provide a convincing comparison of both manual annotation and "CellReg" to demonstrate the tracking performance of "Track2p". Applying this tracking algorithm, Majnik and colleagues characterized hallmark developmental changes in spontaneous network activity, highlighting the impact of longitudinal imaging approaches in developmental neuroscience. Additionally, the code is available on GitHub, along with helpful documentation, which will facilitate accessibility and usability by other researchers.
Weaknesses:
(1) The main critique of the "Track2p" package is that, in its current implementation, it is dependent on the outputs of "Suite2p". This limits adoption by researchers who use alternative pipelines or custom code. One potential solution would be to generalize the accepted inputs beyond the fixed format of "Suite2p", for instance, by accepting NumPy arrays (e.g., ROIs, deltaF/F traces, images, etc.) from files generated by other software. Otherwise, the tool may remain more of a useful add-on to "Suite2p" (see https://github.com/MouseLand/suite2p/issues/933) rather than a fully standalone tool.
(2) Further benchmarking would strengthen the validation of "Track2p", particularly against "CaIMaN" (Giovannucci et al., eLife, 2019), which is widely used in the field and implements a distinct registration approach.
(3) The authors might also consider evaluating performance using non-consecutive recordings (e.g., alternate days or only three time points across the week) to demonstrate utility in other experimental designs.
-
Reviewer #3 (Public review):
Summary:
In this manuscript, Majnik et al. developed a computational algorithm to track individual developing interneurons in the rodent cortex at postnatal stages. Considerable development in cortical networks takes place during the first postnatal weeks; however, tools to study them longitudinally at a single-cell level are scarce. This paper provides a valuable approach to study both single-cell dynamics across days and state-driven network changes. The authors used Gad67Cre mice together with virally introduced TdTom to track interneurons based on their anatomical location in the FOV and AAVSynGCaMP8m to follow their activity across the second postnatal week, a period during which the cortex is known to undergo marked decorrelation in spontaneous activity. Using Track2P, the authors show the feasibility …
Reviewer #3 (Public review):
Summary:
In this manuscript, Majnik et al. developed a computational algorithm to track individual developing interneurons in the rodent cortex at postnatal stages. Considerable development in cortical networks takes place during the first postnatal weeks; however, tools to study them longitudinally at a single-cell level are scarce. This paper provides a valuable approach to study both single-cell dynamics across days and state-driven network changes. The authors used Gad67Cre mice together with virally introduced TdTom to track interneurons based on their anatomical location in the FOV and AAVSynGCaMP8m to follow their activity across the second postnatal week, a period during which the cortex is known to undergo marked decorrelation in spontaneous activity. Using Track2P, the authors show the feasibility of tracking populations of neurons in the same mice, capturing with their analysis previously described developmental decorrelation and uncovering stable representations of neuronal activity, coincident with the onset of spontaneous active movement. The quality of the imaging data is compelling, and the computational analysis is thorough, providing a widely applicable tool for the analysis of emerging neuronal activity in the cortex. Below are some points for the authors to consider.
Major points:
(1) The authors used 20 neurons to generate a ground truth dataset. The rationale for this sample size is unclear. Figure 1 indicates the capability to track ~728 neurons. A larger ground truth data set will increase the robustness of the conclusions.
(2) It is unclear how movement was scored in the analysis shown in Figure 5A. Was the time that the mouse spent moving scored after visual inspection of the videos? Were whisker and muscle twitches scored as movement, or was movement quantified as the amount of time during which the treadmill was displaced?
(3) The rationale for binning the data analysis in early P11 is unclear. As the authors acknowledged, it is likely that the decoder captured active states from P11 onwards. Because active whisking begins around P14, it is unlikely to drive this change in network dynamics at P11. Does pupil dilation in the pups change during locomotor and resting states? Does the arousal state of the pups abruptly change at P11?
-
-
-
