In my attempt to categorize neurons as "small," "medium," and "large," I have difficulty in doing so objectively, as size runs along a continuum. Where does the cut off from "small' to "medium" begin? What do you do with neurons that fall in an in between?
I am constantly reading the literature that loosely refers to classes of neurons based on size, but have not encountered methodology as to how these neurons were placed in said categories. I would like to avoid falling into circular reasoning and imposing a construct on a biological phenomenon.
We are using a neuronal cytoplasmic marker, and as such, we are limited to measuring neuronal somas, as we cannot visualize dentritic arbor or spine morphology.
Any papers, suggestions or ideas are appreciated in advance.
Hi!
This turned out to be a bit more challenging that I expected. Neurons can be classified in many different ways, but when it comes to soma size, scientists seem to often drop the emphasis on continuum and instead identify unique neuronal types by combinations of features, including soma size and location. Examples of this pattern include the Betz and Purkinje cell. But to label cells this way seems to be exactly what you're trying to avoid - imposing a construct on a biological phenomenon.
Still, if the diameter of a cell you've found approximates that of a cell type generally known as "large", "average" or "small" for it's soma size, you can probably feel safe justifying your categorization by analogy. The soma of a neuron can vary from 4 to 100 micrometers in diameter. Betz cells approach 100 micrometers in size, which is considered "quite large" or "giant". Medium spiny neurons (MSN) are ~15 micrometers in diameter and considered "medium-sized". The cerebellar granule cells measure around 4 micrometers and are considered among the smallest neurons in the nervous system.
I'm not sure if that helped, but you're probably right that that there is no clear standard by which to call cells large or small, which might explain why you haven't seen any in your readings.
Al, Martini, Frederic Et. Anatomy and Physiology' 2007 Ed.2007 Edition. Rex Bookstore, Inc. p. 288. ISBN 978-971-23-4807-5.
Bear, M. F., Connors, B. W., & Paradiso, M. A. (2007). Neuroscience: Exploring the brain. Philadelphia, PA: Lippincott Williams & Wilkins.
Hi!
This turned out to be a bit more challenging that I expected. Neurons can be classified in many different ways, but when it comes to soma size, scientists seem to often drop the emphasis on continuum and instead identify unique neuronal types by combinations of features, including soma size and location. Examples of this pattern include the Betz and Purkinje cell. But to label cells this way seems to be exactly what you're trying to avoid - imposing a construct on a biological phenomenon.
Still, if the diameter of a cell you've found approximates that of a cell type generally known as "large", "average" or "small" for it's soma size, you can probably feel safe justifying your categorization by analogy. The soma of a neuron can vary from 4 to 100 micrometers in diameter. Betz cells approach 100 micrometers in size, which is considered "quite large" or "giant". Medium spiny neurons (MSN) are ~15 micrometers in diameter and considered "medium-sized". The cerebellar granule cells measure around 4 micrometers and are considered among the smallest neurons in the nervous system.
I'm not sure if that helped, but you're probably right that that there is no clear standard by which to call cells large or small, which might explain why you haven't seen any in your readings.
Al, Martini, Frederic Et. Anatomy and Physiology' 2007 Ed.2007 Edition. Rex Bookstore, Inc. p. 288. ISBN 978-971-23-4807-5.
Bear, M. F., Connors, B. W., & Paradiso, M. A. (2007). Neuroscience: Exploring the brain. Philadelphia, PA: Lippincott Williams & Wilkins.
Do you need to have an a priori set of threshold values? You may want to make the measurements first and draw a histogram to see if there are peaks and troughs. Depending on how you are defining size, the histogram may be two-dimensional (lengths of long and short axes, for example) or higher-dimensional. You can choose thresholds to maximize the reproducibility of your categorization.
Jun, I have plotted all my neurons on a scatter plot vertically to observe their distribution. It was still difficult to determine where to draw the cut off line. This is also not the most arbitrary way to go about categorizing. For the large ones, there was noticeable clustering, but I could not see any breaks to determine a threshold for "medium" and "small" neurons.
There is opinion, that in DRG neurons, with soma diameter less, than 25-30 micrometers are nociceptive neurons. Also these cells are thought as "small". More about classification of "small" DRG neurons you can find here:
http://www.ncbi.nlm.nih.gov/pubmed/23376589
http://www.ncbi.nlm.nih.gov/pubmed/24693454
I think you are up against the historical nature of mapping of neurons. Since each area of the brain was characterized at a different time and place, the Relative size of neuron somas as characterized as small medium and large, is changeable and inexact. Relative size depends on the state of microscopy and accuracy of possible measurement as done at the time of characterization of the neurons.
This might be annoying if you are expecting some standard to fit across the brain, but, usual process was to characterize the size both by relative measurement and by size in some set metric as used at the time. If you ignore the less accurate description and use the size you will be a lot better off. However to get this level of accuracy you will have to probably go back to the original documents.
You might want to try a clustering algorithm similar to those used to isolate single units from multi-unit recordings. That would provide you with an objective way of classifying your cells. Alternatively, instead of establishing somewhat ambiguous criteria of small, medium and large, you may want to consider the size distribution as a whole and ask whether the distribution is different under different conditions. For an example see a paper by Fu et al., in JCI 2014 about the role of retinoblastoma gene in enteric nervous system development. There is a cumulative frequency distribution curve in that paper that nicely outlines the difference in the populations.
Rather than looking at histograms it may be more fruitful to use a clustering algorithm, or a decision tree. See for example, the R packages k-means and tree respectively.
If you are having trouble classifying objectively based on relative criteria, why do it?
Is it totally necessary? Especially since, as you said, its a continuum. I mean, this is just my opinion but it would seem more informative to classify a cell type based on something like a cell-specific marker or functional role.
Thank you all kindly for your suggestions and references.
Edward, it is necessary to my research. I am looking for strategies to work within these constraints. We will most definitely employ other IHC methods instead of simply labeling neuronal somas. Shape and size is a component of cell structure and function--and as such, it needs to be addressed accordingly. We are testing the hypothesis that statins that cross the blood brain barrier affect neuron shape and size. I am specifically examining neurons in the song motor pathway. A detailed anatomical paper by Fortune and Margoliash (1995) references large, medium and small neurons. In my attempt to use their paper as a basis for my research, I came across this problem. Many of these suggestions others have provided above have limitations. In any case, if I used a cluster analysis to establish a threshold for size categorization, I have to acknowledge the limitations and discuss them in detail in the discussion section of our manuscript.
Ah I see, that makes sense. That sounds super interesting, best of luck with that!
Please have a look on this paper:
Behzadi, G., Ganji, F., 2005. Morphological alteration in oro-facial CGRP containing motoneurons due to congenital thyroid hypofunction. Peptides 26, 1486–1491
Could you indicate the cytoplasmic marker has been used? I believe the real size of neurons differs upon the chemical staining; Nissl staining, enzyme histochemistery, immunostaining or different retrogradely tract tracing substances. Interesting to hear the other opinions about. Good continuation for you.
Gila, thanks for reference. We are using the antibody to Hu, a protein expressed in neuronal cytoplasm. It labels neurons as young as two hours.
Very interesting Alicia. By the way, certainly you are using an image analyzer software?
Nice question, however, neurons are mostly classified either upon their structure (depending on the number and distribution of dendrites) or their function (sensory, motor, interneurons, etc.).
Also, neurons are classified according to the type of neurotransmitter they produce (adrenergic, cholinergic, somatostatinergic, gabaminergic, etc.).
The giant neurons are mostly present in invertebrate animals like slugs and annelids.
Good luck in your work.
DRG neurons are classically characterised by size. Numerous publications over the years have used a "binning" approach to size classification of DRG neurons, where DRG neurons are binned into a diameter range that covers 10-20 um. Data can then be graphed as % of neurons on the Y axis and increasing diameter range on the x-axis. This is a particularly useful approach for comparison of DRG neurons between WT mice and transgenic mice.
Measuring cell bodies in sections is a pain in the neck. Plagued by many constraints to be factored in (territory, age, model, population...) and artifacts to be accounted for (fixation, embedding, sectioning, staining...). Only if you think the expected data are worth it, do it. But do it in a proper manner. If you are nor familiar yet with stereology, check any basic book on that (Howard and Reed, or West, for example). If you are, but need more specific information on specific details, let me know.
If there is enough demand for it, you could automate the actual measuring step and limit yourself to graphing the results according to territory, population etc.
serial sectioning and morphometric study ,stereological technique mey be availablen for your for your demand best whishes
- Badges
- Science topic
- Similar topics
- Biological Science
- Cell Biology