Distinct In Vivo Electrophysiological Profiles of Mediodorsal Thalamus Subdivisions

Curation statements for this article:
  • Curated by eLife

    eLife logo

    eLife Assessment

    In this work, Panti et al. examine the electrophysiological properties of mediodorsal thalamus subdivisions in mice. This study is important for understanding how this region may support multiple integrative functions. However, the claim that the central portion of the mediodorsal nucleus is physiologically distinct is currently supported only by incomplete evidence.

This article has been Reviewed by the following groups

Read the full article

Abstract

The mediodorsal nucleus of the thalamus (MD) plays a key role in complex cognitive processes, and its dysfunction is linked to various neurological and psychiatric conditions. The MD is divided into three parts that vary in anatomical connectivity, molecular expression, and ex vivo electrophysiology. Here, we describe in vivo single-neuron electrophysiological recordings across the three MD subdivisions in head-fixed, behaving, adult wildtype mice. We report large differences in extracellular waveforms, spiking activity, and burst firing characteristics across MD subdivisions. Specifically, central MD subdivision (MDc) neurons showed markedly increased waveform amplitude, higher spontaneous firing rate, and increased burst firing compared to medial (MDm) and lateral (MDl) subdivision neurons. Single-neuron electrophysiology features were sufficient to classify neurons into respective anatomical subdivisions. Hierarchical clustering revealed MDc is distinct from MDl and MDm and more akin to other, non-MD thalamic nuclei. Together, these data suggest distinct in vivo electrophysiological profiles of MD subdivisions, posing implications for investigating MD thalamus in health and disease.

Article activity feed

  1. eLife Assessment

    In this work, Panti et al. examine the electrophysiological properties of mediodorsal thalamus subdivisions in mice. This study is important for understanding how this region may support multiple integrative functions. However, the claim that the central portion of the mediodorsal nucleus is physiologically distinct is currently supported only by incomplete evidence.

  2. Reviewer #1 (Public review):

    Summary

    In this paper, Panti, Stagg, Sharott & Stedehouder investigate the activity of the mediodorsal thalamic nucleus (MD). In rodents, it has long been known that the MD is divided into three distinct subregions (central, medial, and lateral), differentiated by their afferent and efferent connections. However, until now, no one had demonstrated distinct electrophysiological properties in vivo across the three MD subdivisions.

    In this study, the authors recorded, in head-fixed mice performing a sensorimotor task, the extracellular activity of neurons from the different MD subdivisions using Neuropixels probes. They revealed differences in extracellular spike waveform (amplitude, repolarization latency, peak-to-trough ratio) as well as spiking and burst activity, with MDc neurons exhibiting higher firing rates, lower rate variability and higher burst index and rate than the other two subdivisions. They also show that a classifier based on these parameters allows categorization of neurons according to their MD subdivision, in particular the MDc.

    Strengths

    There is growing interest in the role of the mediodorsal thalamus, and its characterization is very important. MD subdivisions, due to their specific anatomical connectivity, can support distinct functions. The question of identifying MD subdivisions based on extracellularly recorded spikes is particularly compelling.

    The dataset is very strong and large and is well described in the supplementary figures. The work is technically sound, and every step of the methods used is well presented. The authors present all results from the different cohorts honestly, and they also perform various controls.

    Weaknesses

    I have concerns regarding the meaningfulness of the waveform analysis and statistical power, as well as the interpretation of the firing activity of MD units.

    (1) Waveform analysis and statistical interpretation:

    a) In Figure 2, the distinction between waveforms across subdivisions is unclear (especially Figure 2B). The within-subdivision variability appears far higher than the between-subdivision differences, raising questions about the meaningfulness of these statistical differences.

    In extracellular recordings, spike shape differences are inherently noisy. Even when comparing extremely distinct cell types, such as pyramidal cells vs. interneurons, clustering can still show at least some partial overlap, despite the major morphological differences in spike waveforms. The effect sizes shown here suggest that the observed statistical differences may potentially lack biological significance. I thus suggest trying alternative statistical tests that might be more robust, for example, considering the animal and session as random effects with a mixed-effects model/regression. Importantly, this approach can be applied to firing rate metrics (even though the effect sizes are larger, especially for the firing rate).

    b) For a single unit, how does spike waveform evolve over the course of a recording session? In extracellular recordings, amplitude drift for the same cell can happen. Can the authors describe how they control for/account for that in their data curation?

    (2) Data interpretation of spiking metrics:

    a) The MD activity was recorded during a sensorimotor task (auditory cues, reward delivery, lever presses). The authors mention, "Across all cohorts, recordings were obtained acutely from head-fixed behaving mice performing a sensorimotor task over several days" (results, lines 98-99). It is unclear which periods the authors analyze. For example, the authors mention "spontaneous firing rate over the entire recording window" (line 164). This is confusing because the authors also mention, "These large firing rate differences across MD subdivisions were similarly evident in passive interval windows (periods outside of bar movement or auditory stimulation related to the behavioral task" (lines 168-170). It is thus unclear what is meant by spontaneous activity.

    b) Regarding the results, the differences the authors observed between MD sub-divisions in terms of firing rate and burst characteristics may reflect the engagement of those different subregions to different aspects of the tasks. Indeed, it has long been known that there are differences in anatomical connectivity that might explain the differences observed. The activity is neither naïve nor spontaneous but rather context-dependent. The data set is very strong, with many MD units recorded across different subdivisions during such tasks. Therefore, analyzing MD units in regard to these different behavioral events is critical to interpret the results on firing rate and burst activity. For example, MDc might only present strong burst activity, higher firing when the animals are pressing the bar or, in general, when they are strongly engaged in the task (which can thus also be seen during the passive interval).

    (3) How correlated is the activity of MD units recorded simultaneously within the same subdivision? Are there synchronized firing patterns such as assemblies? If so, does this vary as a function of MD subdivisions?

    (4) An accuracy of 52.5% of a classifier compared to an effect of 33% chance level is not strong but rather moderate. Most importantly, there is spatial autocorrelation. From my understanding, the authors used a ten-fold cross-validation with 90% training and 10% testing. Was the cross-validation spatially blocked (by MD subdivision)? If not, I would strongly suggest doing so.

  3. Reviewer #2 (Public review):

    Summary:

    The authors perform extracellular Neuropixels recordings in awake mice to characterize the firing properties of neurons in the three major subdivisions of the mediodorsal nucleus of the thalamus, which plays roles in executive functioning and social behavior.

    The authors found that the medial, central, and lateral subdivisions did have substantial differences in their firing properties that allowed for estimation of subdivision better than chance.

    Strengths:

    The strengths of the manuscript include the importance of this topic, which expands our understanding of MD thalamus (1) so we can move beyond generalizing findings from sensory thalamus to higher-order thalamus and (2) to guide future neurostimulation strategies. The strengths include the size of the data sets and the number of animals sampled. The manuscript is well-written, and the figures are clear. The results support the conclusions. This data set will be valuable for scientists involved in studying the higher-order thalamus and lower-order thalamus.

    Weaknesses:

    The major suggestion I have is to increase the rigor by using a linear mixed-effects model to account for the effects of cohort and animal. I don't think this additional analysis will change the findings or conclusions, but it will strengthen the statistical foundation of the paper. Studying only male mice limits the generalizability of the findings.

  4. Reviewer #3 (Public review):

    The role of the mediodorsal thalamus (MD) in cognitive functions is well-established in humans as well as in several primate and rodent animal models. However, MD is not a homogeneous structure but is composed of at least three parts, the medial, central, and lateral (MDm, MDc, MDl) subdivision that bear distinct molecular profiles, input-output organization and functions. Whether similar heterogeneity exists in in vivo electrophysiological features among MD subdivisions has been unknown until now. Panti and colleagues used high-density electrophysiological measurements from distinct parts of MD in head-fixed, behaving mice to compare their basic characteristics, such as firing rate, waveform, and burstiness. They found that, in general, MDc neurons are markedly distinct from their neighbors, while MDm and MDl cells rather share similar profiles.

    This paper has a gap-filling nature, indeed. However, there are some issues that need to be clarified. Please see my detailed comments below.

    (1) It is not clear whether the observed differences between MD subdivisions are already present in the baseline or triggered by the behavioral paradigm.

    It is mentioned in the abstract and introduction that all the recordings were obtained during behavior tasks. There is a detailed description of these tasks in the methods section. However, there is not a single behavior-associated data point in the paper. If the latter is the case, task-specific analysis is necessary since any electrophysiological profiles can be influenced by the presented stimuli and/or by the triggered behaviors, by cue presentation, delayed period, reward, hit or false trials, motor (licking) actions, etc.

    (2) There were several recordings from the same mice. Even if the probes were stained with DI, how could the individual trajectories be separated after many days? And how could the individual cells be matched with a recording day and a thalamic subdivision? Even if the individual penetrations were separated by 2-300 um, it is likely that these were finally overlapped. As it is mentioned and widely accepted that units can be recorded from a 100-150 um distance, it is likely that the same cells were monitored in different recording sessions. How do the authors compensate for it? Were each recorded analyzed separately and then pooled?

    (3) Going further on this line, every penetration causes damage in the tissue. To rule out the possibility that the damage differently affected the recordings from distinct thalamic territories, each session should be individually handled and then compared before any pooling.

    (4) MD subdivisions were classified using atlases. However, it is difficult to separate MDm and MDL territories from the neighboring paraventricular (PVT) and intralaminar (IL) thalamic regions (respectively) due to the lack of sharp borders. (Furthermore, some literature data suggest the existence of a transition zone between PVT and MDm, which makes it even more difficult to separate these two thalamic areas without a cell-type-specific analysis.) It could be a possibility that MD is homogeneous; however, MDm and MDl differ in the presented dataset due to their 'contamination' with PVT and IL cells. Although the statistical powers (P values) are high, the actual differences are around 10% in most cases. Furthermore, the range of the data points is similar, only their distributions show differences. Thus, a small portion of cells with distinct kinetics could easily shift the population result.

    In addition, a cell can be recorded from 100-150 um, which is the dimension of the MD subdivisions. Thus, it is extremely difficult to identify the exact location of a cell recorded in the medial or lateral parts of the MD. Excluding those cells that are closer than 100 um may not solve this situation because outlining any brain section with a corresponding atlas can not provide a sufficient spatial resolution. Sections with DI tracks could be counterstained with neurochemical markers (visualizing thalamic subdivisions) could also help the issue.

    Another, complementary solution could be the comparison of MDm and MDl cells with those that are presumably located in the PVT and IL (respectively).

  5. Author response:

    We thank the Editors and the Reviewers for their evaluation of our work and constructive feedback. We appreciate the recognition of the importance of our study, the strength of our dataset, and its contribution to understanding the electrophysiological organisation of the mediodorsal thalamus (MD). The main concerns raised by the Reviewers include the influence of behavioural and experimental events on examined physiological metrics, the uncertainty of anatomical localisation of MD neurons, and the statistical framework employed. We think these concerns can be addressed through both a clarification of our approach and through additional analyses, and we would like to provisionally outline how we plan to address them in a revised manuscript.

    First, we will more thoroughly untangle the influence of various behavioural events on the electrophysiological differences across MD subdivisions. We will describe more clearly how we calculated the physiological metrics used throughout the manuscript. Furthermore, we will extend our current analyses to examine in more detail how physiological characteristics throughout MD subdivisions are influenced by task-related events and animal engagement.

    Second, we will more extensively describe our anatomical localisation approach and past and current considerations for neuron assignment. We will extend our considerations for both the lateral, medial and ventral borders of the MD in general as well as its parcellation into subdivisions. We will provide additional data to support our claim of accurate localisation of neurons, as well as data to suggest minimal effects of contamination of non-MD thalamus neurons in MDl and MDm.

    Lastly, we will strengthen the statistical framework of our study. We will implement mixed-effects models to account for the hierarchical structure of our dataset and provide a more rigorous statistical foundation of the findings.

    Overall, we believe the recommendations and planned revisions will address the concerns raised by the Reviewers and strengthen the interpretation of our findings. We thank the Reviewers for their constructive comments and look forward to incorporating their suggestions into the revised version of our manuscript.

  6. 1/ NEW PREPRINT: The mediodorsal thalamus (MD) is critical for cognition and implicated in several brain disorders. Panti et al. reveal that the MD is much more diverse than previously thought. doi.org/10.64898/202...

  7. 1/ Preprint from Brain Network Dynamics Unit: The mediodorsal thalamus (MD) is critical for cognition and implicated in several brain disorders. Panti et al. reveal that the MD is much more diverse than previously thought. doi.org/10.64898/202...

  8. Distinct In Vivo Electrophysiological Profiles of Mediodorsal Thalamus Subdivisions https://www.biorxiv.org/content/10.64898/2026.09.02.748524v1

  9. Distinct In Vivo Electrophysiological Profiles of Mediodorsal Thalamus Subdivisions https://www.biorxiv.org/content/10.64898/2026.09.02.748524v1