Tuesday, May 15, 2012

Dendrites of Direction

Neurons in the Retina (Figure 1h Kim et al., 2008)
This is an image of a piece of retina with neurons labeled with a specific marker (JAM-B). Notice something about them? Yep, their dendrites are all pointing downwards. If you are regular Cellular Scale reader, you will remember that dendrites take on many different shapes, and that these shapes often mean something with regard to the function of that cell. 

A 2008 paper shows that this downward direction in these dendrites is no coincidence. These neurons are sensitive to visual input moving in a specific direction, the same direction that their dendrites are pointing.  In other words, when a visual stimulation (such as a a bar or dot) is moving across the retina in the soma to dendrite direction, these neurons are most active.  When the visual stimulation moves in the opposite direction, these neurons are the least active. 

Figure 2e, Kim et al., 2008

This diagram shows the direction of the dendrites (green line), and the direction of movement which activates that neuron the most strongly (red line).  This is just one example, but on average the dendrite direction and the preferred stimulus direction matched up for these neurons. 
Because the lens of the eye functionally reverses the visual world, this means that since the dendrites of these neurons point down, the actually respond to upward motion.
A direct match between the shape of the dendritic tree and the function of these neurons is a huge step toward understanding the way that Form and Function influence each other in the brain. The authors end this paper by asking

"One outstanding question is why the mouse has invested so heavily in sensitivity to upward motion." (Kim et al., 2008)

This could lead to speculation on mouse evolution and why a mouse would need to be extra-sensitive to upward visual input. But I think that is a goose chase, just because there are cells in the retina whose form and function match nicely for upward motion, doesn't mean that mice are actually more sensitive to upward motion. (A motion detection behavioral test is necessary to make that claim, and I don't know of any done on mice)

In fact, the same group more recently (Kay et al., 2011) found that there are four similar classes of cell responsive to each of the four cardinal directions.  These cells have some dendrite-direction correlation, but it is not as strong and clear cut as the upward sensitive cells specified in the 2008 paper. 

What is particularly interesting is that, while the (J-RGCs) cells in the 2008 paper have such strong correlations between dendrite direction and stimulus direction sensitivity, the cells (BD-RGCs) described in the 2011 paper do not.  From the discussion:

"The correspondence of dendritic asymmetry with preferred movement direction in BD-RGCs resembles that in J-RGCs, a far more strikingly asymmetric group of OFF-DSGCs that we described recently (Kim et al., 2008, 2010). We suspect, however, that the association differs in the two cases. Both J- and BD-RGCs include some cells whose arbors appear symmetric. The symmetric J-RGCs are not direction selective, supporting the idea that structure underlies function for these cells (Kim et al., 2008). In
contrast, structurally symmetric BD-RGCs are as direction selective as asymmetric ones, suggesting that for these cells structural asymmetry does not determine directional preference." (Kay et al., 2011)

In other words: Some cells are direction-sensitive without their dendrites being weighted to one side. 

So the really exciting questions are: What are the molecular and cellular mechanisms that make these cells directionally sensitive, and is the dendritic orientation necessary for direction sensitivity?  If an upward motion cell was somehow transplanted in the opposite orientation, would it become a downward motion cell?

I suspect that just as computational neuroscience helped us understand the dendrite-based frequency sensitivity in the bird brain, a computational model would help us understand how a cell could respond maximally to soma-dendrite directional motion.  

© TheCellularScale

ResearchBlogging.orgKim IJ, Zhang Y, Yamagata M, Meister M, & Sanes JR (2008). Molecular identification of a retinal cell type that responds to upward motion. Nature, 452 (7186), 478-82 PMID: 18368118

Kay JN, De la Huerta I, Kim IJ, Zhang Y, Yamagata M, Chu MW, Meister M, & Sanes JR (2011). Retinal ganglion cells with distinct directional preferences differ in molecular identity, structure, and central projections. The Journal of neuroscience : the official journal of the Society for Neuroscience, 31 (21), 7753-62 PMID: 21613488


  1. I hope the direction of the dendrites IS important for the motion sensitivity--that'd be neat. I recall Martha Livingstone putting forth an interesting suggestion about how the architecture of a cell could be related to the functional motion sensitivity. I think the basic idea was that connections from a retinotopic area were lined up along a dendrite such that motion in the correct direction would result in a summation of graded potentials that wouldn't occur if the motion was in the other direction. This also depends upon speed of conductivity, I suppose.

    Anyway, I don't think her theory stood up to scrutiny, but I remember feeling that it was a really elegant way of doing things and hoping that the brain really did work that way.

  2. R, are you referring to this paper?
    If so, it's really interesting! Figure 8, shows 3 ways that the specific receptive field of a visual (cortical) neuron could become a certain shape, and way number 3 has an interesting dendritic map.

    She's talking about the visual cortex as far as I can tell, while the cells in this paper are retinal ganglion cells. So she's not talking about the same thing exactly, but is definitely addressing the form and function question.

  3. I thought that the direction selectivity had to do with asymmetrical inhibitory innervation of RGCs by amacrine cells?

  4. These J-RGCs are only a small subset of RGCs, so they might not follow the general pattern. (The JAM-B labeling only labels a few hundred cells in the retina.)
    Also, the authors suggest the specific location of these dendrites restrict the amacrine inputs onto them:

    "Dendrites of J-RGCs arborized in a narrow band between processes of dopaminergic and cholinergic amacrines13, indicating a sublaminar restriction (Fig. 1f) and suggesting that J-RGCs receive few synapses from these two amacrine classes. In contrast, processes of CD15-positive and AII amacrines did overlap with dendrites of J-RGCs (Supplementary Fig. 3) and may therefore provide input to them."

    I am not a retina expert by any means, so here's some wild and uninformed speculation:
    Maybe most other direction sensitive RGCs don't need asymmetrical dendrites because they have asymmetrical amacrine innervation, but these ones need lopsided dendrites to compensate for a (possible) lack of lopsided amacrine input.

  5. Ah, I hadn't seen that paper, but it seems similar to the ones I remember. I read them years ago now when I was doing a bit of work with Donald Mitchell and Kevin Duffy. Vision is outside of my area, too--I have more experience in auditory psychophysics and animal learning, and don't know as much about the nittygritty neuroscience stuff as I'd like--but I really enjoy the work of researchers who relate neuronal function to neuronal architecture. It seems elegant to me.

    In any case, I'm definitely enjoying your blog. :D

  6. Do you normally compose only for this domain or maybe for any other Internet portals?

  7. I normally only write here at The Cellular Scale. Why do you ask?