2nd pass review, 5/6/09 by Mel
1. Regarding this sentence: "Pyramidal neuron dendrites may also form synapses onto other neurons or release retrograde signaling molecules (e.g. endocannabinoids), so communication may be bidirectional."
Comment: I have never heard of this; it's possible the way it is worded that the reader will overgeneralize and believe that output synapses are general features of pyramidal neuron dendrites. (Are they? Where does this occur?)
2. Fix grammatical error here: "The extensive branching of the dendrites and the axon branch allows a single to neuron to communicate with thousands of other neurons in a network".
3. Regarding: "Pyramidal neurons receive synaptic inputs from tens of thousands of excitatory synapses and several thousand inhibitory synapses", this differs from the sentence above saying it's typically 10,000. I also again refer to one of the most studied areas in the cortex V1; many pyramidal neurons there according to Elston have 2-3K synapses in their basal dendrites. It all depends on the area and layer.
4. Should add (Gordon, Polsky, Schiller 2006) to this list: "...as well as a signal for dendritic plasticity (Golding et al. 2002; Kampa et al. 2007; Remy and Spruston, 2007; Losonczy et al. 2008). "
5. "coincidence detection occurs when separate excitatory inputs activated the tuft": sounds like should be "activate the tuft".
6. "originating in forebrain nuclei": would be helpful to have a references.
7. Figure 9 caption should mention where those neurons are (entorhinal cortex?)
8. In general: Too bad the references are links to pubmed abstracts. That would be really helpful. I kept wanting to click on them to see what they were.
9. This sentence should refer to Figure 12: "In keeping with this, pyramidal neurons contain a subset of large, structurally specialized synapses, which may result from learning and LTP (Geinisman, 2000). "
Original review by Mel
A clear, well written short but comprehensive review. A number of small improvements are possible, however.
1. Basal dendrites don't necessarily emerge from the base of the soma, at least in neocortex. I'm not an expert on this, but neocortical basal dendrites appear to be more stellate in form. If you look at Guy Elston's papers (e.g. Elston and Rosa, 1998), the basal trees of layer 2-3 pyramidal cells appear almost perfectly spherical. In keeping with this, spiny stellate cells are sometimes thought of as apical dendrite-less pyramidal neurons. It may be worth raising the point that pyramidal neurons are not all the same very early in the discussion, even before talking about their basic morphological properties.
I think "basal" dendrites are so named because they emerge from the "base" of the pyramid, while "apical" dendrites emerge from the "apex" of the pyramid.
2. Say what "principal cell type" means, e.g. add something like, "making up 80% of the total neuronal population, the rest being made up of inhibitory interneurons". Otherwise, principal can sound like "most important".
3. Ironically, the word "dendrite" is unclear in meaning. In the expression "..have multiple dendrites", it would be good to specify what one dendrite is supposed to be, vs. a tree vs. a subtree. When one says "has multiple dendrites" one usually means branches leaving the cell body. But then, that branch branches, so where does "the" dendrite end? To me, a single dendrite semantically would not include branches; that would be a tree or subtree. Of course, the whole story is contained in the dendrogram. But I think the standard that should be applied is that it should be possible to picture any term used, and right now, I can't picture what "dendrite" refers to.
I think the meaning of the word "dendrite" is pretty clear. A dendrite can branch, as can an axon. Neurons typically have one axon and multiple dendrites. This refers to the number of processes emerging from the soma.
4. "each basal dendrite branches several times": may be better to say , "up to several times" since some don't branch or branch once.
5. "several oblique dendrites": In CA1 it can be up to 20 or 30; Is that still several?
Yes, I think this is the best word here. I think to say "many" would be an overstatement for some pyramidal neurons. To say a "few" would not be enough for some pyramidal neurons. "Several" falls in the middle of these two. I'm open to suggestions, but I think it's ok as is.
6. Where you say it's an oversimplification that for pyramidal neurons, dendrites=input and axon=output, it might be good to put a parenthetic comment about what the more complicated reality is. Do you have in mind diffusible retrograde factors, gap junctions, or actual vesicular release from pyramidal neuron dendrites? I think one should not allude to complexity without giving any indication as to what it is. Either elaborate a bit, or I suggest drop the elaboration.
7. The phrase: "Both the dendrites and the axon branch extensively" appears twice in close succession.
8. Should say somewhere about how many synapses there are on a pyramidal neuron. But this raises another issue: there is no indication as to the range of sizes and synapse counts of different pyramidal neurons. The above mentioned Elston and Rosa paper reports that there are 13 times more spines on an IT neuron than a V1 neuron. I think it's 17 times more for a prefrontal neuron. That's worth mentioning somewhere lest the reader conclude (as suggested by the picture in this article) that all pyramidal neurons are about the same size.
I added a sentence about the variation in size. The wording regarding the number of spines and synapses does not give a precise number, so I think it's ok for an encyclopedia type of article.
9. There is a peculiar lack of mention of NMDA channels in discussion of dendritic spikes; these are main current carriers in basal dendritic spikes in neocortex, at least. Also, in the section on synaptic integration, it seems a sentence should be added alluding to individual thin apical oblique and basal dendrites as integrative compartments, within which mixed NMDA/sodium/calcium spikes can be driven by spatio-tempoally concentrated inputs (e.g. Mel, 1993; Poirazi et al. 2003; Ariav et al. 2003; Polsky et al. 2004, Milojkovic et al. 2004, 2005; Nevian et al. 2006; and Losonczy and Magee 2006, 2008).
I made some revisions along these lines.
Thanks for the helpful comments.
Comment on missing content
In the section Dendritic excitability the text satrts mid sentence indicating that perhaps some text is missing.
that gives rise to its computational power. Much of this integration occurs in the dendrites, so it is sometimes referred to as dendritic integration (Häusser et al., 2000;
Additional minor suggestions (16.4.2013, by Dan) :
1. Under Structure section - "...the axon allows a single to neuron to communicate with thousands of other neurons in a network."
Should the first to be omitted?
2. Nelson Spruston’s website link is broken, should it be redirected to his current website ]?