JOURNAL CLUB: Somatosensory processing in mouse sensory and motor cortex
Eric Thomson
Tuesday, 29 January 2008 20:09 UTC
Introduction
If you watch a rat or mouse running about, one thing that will stick out is how frequently they use their whiskers to actively explore the world. Rodents’ long facial whiskers provide them with surprisingly keen sensory abilities. For instance, their ability to discriminate textures rivals that of humans using their fingertips. The importance of the whiskers in the rodent somatosensory system is reflected in its cortical representation: almost half of primary somatosensory cortex (S1) in rodents is occupied by neurons that respond primarily to whisker stimulation.
In layer IV of S1, each whisker is represented by a ‘barrel’, a 500 micron-wide circular region of cortex. The stereotyped and relatively simple arrangement of the whiskers and barrels have made the whisker system a popular and fruitful model system for studying the cortical processing of sensory information.
Methods and results
Spatiotemporal Dynamics of Cortical Sensorimotor Integration in Behaving Mice, by Ferezou et al. (paper available here), is an experimental tour de force. They applied a voltage-sensitive dye (VSD) to the cortex of head-fixed mice to measure subthreshold voltage changes in the cortical surface (layers II/III). They analyzed the activity in a large area of cortex that included S1 and primary motor cortex (M1). This allowed the authors to examine sensory responses in S1, M1, and the relationship between responses in these two areas.
In the first set of experiments, they stimulated a single whisker in anesthetized mice. The earliest cortical response was localized to the corresponding barrel in S1. Surprisingly, within 8 ms this localized response spread to the entire whisker representation in S1 as well as M1. Interestingly, they found that different regions of M1 were activated by different whiskers: that is, there is a map of the whiskers in M1. In addition, S1 activation and inactivation experiments showed that that S1 activity was both necessary and sufficient for stimulus-dependent responses in M1.
The second set of experiments was performed in awake, head-fixed mice. When a mouse moved a whisker into an obstacle, the same pattern of cortical activation as described above was observed (i.e., localized S1 activation followed by broad S1 activity and M1 activity). Indeed, Figure 7B shows a trial in which most of the cortical hemisphere was activated within 40 ms of the touch.
Finally, the authors demonstrated that the sensory response is strongly influenced by the general behavior of the mouse during whisker stimulation. For example, when the mouse was initially resting but started to whisk in response to whisker stimulation, the same pattern of activation described above was seen. However, different responses were observed when the stimulation did not evoke whisking behavior in resting mice or when the stimulus was delivered in the middle of a whisk cycle.
Discussion
Ferezou et al. showed that subthreshold responses to whisker stimulation can be quite broadly distributed, often extending into M1. This suggests that M1 does not have a purely motor function, but serves also to process sensory information. While M1 projects directly to the brain stem and spinal cord to coordinate motor activity, its tight link with S1 opens up interesting questions about its role in sensory processing and sensorimotor transformations. Also, the sensory response in S1 and M1 depends on the behavioral state of the animal, suggesting that sensory processing isn’t a stationary process, but is sensitive to the context in which a stimulus is delivered. So, when someone asks how a mouse’s cortex would respond to a given stimulus, you probably have to ask, “What is the critter doing?”
Discussion questions
Note: some of these are mentioned in the paper.
1. How broadly does the sensory response spread? Does it spread to visual, auditory, and other sensory cortices (and vice versa)? What implications would such broad sensory effects have for theories of sensory processing in more ethologically realistic contexts when inputs are coming in from multiple modalities?
2. In nature, mice explore objects with many whiskers simultaneously. Would you expect the results in the paper to generalize to more naturalistic stimuli?
3. What are some possible functions of the observed differences in the sensory response when the mouse is in different behavioral states?
4. When the awake mouse was whisking and contacted an object, the typical S1→M1 pattern was observed, but when the whisker was stimulated magnetically during a bout of whisking (the final set of experiments mentioned above), the response looked quite different. What might explain this seeming contradiction in their results?
5. Rodents can still whisk when their cortex is removed (there seems to be a brain stem CPG responsible for whisker movement). What types of control does M1 exert over whisker movements in intact awake rodents? What role does S1 play in such motor control?
6. Rodents have fewer somatosensory areas than primates. Would we see the same sensory influences in M1 of primates?
Updated 30 January 2008 17:50 UTC
-
Replies
-
Dear Eric,
I’m glade someone introduced and reported on this study, because I actually like it a lot.
What would be your answers to the questions?In general I think the connection between S1 and M1 is no suprise, because a cortical area has to coordinate the motor output following sensory stimuli. The results reveal a quite realistic scheme in various condition, especially that cortical responses differ between active and passive whisking!
A major drawback of the study is that to date no one is able to image deeper cortical layers, which would be the most intersting, actually. E.g. deKock and Sakmann show that the major cortical output arises from layer 5b pyramids with latencies comparable to layer 4 and upper layers show low AP responses onto a single whisker deflection. But in this study there is no behaviour…
I am very intersted what subsequent studies will show! -
Mike: I wasn’t too surprised by the M1 sensory response, but for different reasons than you mention.
As I allude to in question 6, since the mouse has relatively few somatosensory areas it wouldn’t be surprising if the sensorimotor transformation had to be compressed. But in primate I’d be a bit more surprised, as there are plenty of sensory areas, premotor areas, etc. to preprocess the sensory information. The areas might be more specialized in the primate. Obviously this is speculative, and someone out there surely knows the primate somatosensory system, and whether there are sensory responses in primate M1, better than I. I haven’t found, and can’t recall seeing, similar results in the monkey.
Also, M1 sensory responses have been observed previously in the rat. In my summary I didn’t mention which results were new and which were extensions of previous work, as it became cumbersome, and they were pretty clear about this in the paper.
What surprised me most in the paper was the incredibly widespread activation that they observed in the awake whisking mouse (e.g., Figure 7B, which if a typical response is just amazing). This is something I have never seen before which could have far-reaching implications.
A nay-sayer might say “Oh, it is just subthreshold activity.” To them I’d point out that it is very unlikely that you would see such long-distance influence from point A to point B without spiking at point A. On the other hand, instead of indicating a causal connection between A and B, such widespread firing could be due to a subcortical structure that is broadcasting some excitatory signal to cortex or entraining the brain to some rhythm. However, they established this was not the case for the S1—>M1 connection, which seems to be a monosynaptic pathway.
-
Eric wrote: “When the awake mouse was whisking and contacted an object, the typical S1→M1 pattern was observed, but when the whisker was stimulated magnetically during a bout of whisking (the final set of experiments mentioned above), the response looked quite different. What might explain this seeming contradiction in their results?”
One possible explanation is that in the normal case there was an attentional mechanism anticipating the perception, as the alpha desynchronization described in a recent Nature Precedings paper by Capotosto et al
-
Alfredo: very interesting—Imagine feeling around in the dark with your hands—attention is usually focused on the somatosensory inputs. Do rodents ever whisk without attending to whisker inputs? (Note I think that whisking isn’t necessary for attention: in the whisker-dependent task I run my rats on they usually don’t whisk.)
Also, one key fact I didn’t mention in the summary was that the stimuli were different in the two cases. In the ‘whisker contacts object’ setup, the whisker movement was halted mid-whisk so to speak when the whisker contacted the obstacle (and the obstacle supplied a force that was along the rostro-caudal axis as whisker movement), while in the second set of experiments the whisker kept moving along the rostral-caudal axis, as it was stimulated using a magnetic field along the dorsal-ventral axis (the whisker had a metallic element attached) orthogonal to the direction of whisker movement. Such stimuli could have quite different effects.
-
I agree with Eric that it is most likely correct that whisking and attention are not linked 100% of the time (as is indirectly concluded in the paper, as demonstrated by the last set of experiments. The attentional state probably modulates the whisking behavior, perhaps after M1 processes S1 information.
Since rodents are so dependent on whisker information, the connection between S1 and M1 seems important to decrease the delay between gathering information and initiating a response, whether it be to whisk more (continue gathering more information) or to follow some other motor-based behavioral pattern.
-
Dear Eric and Noah:
Please check the Abstract below (seems to be good old science). Could the tremor be related to attention instead of Parkinson?
Thanks
Alfredo
Brain Res. 1980 Aug 18;195(2):281-98.
Synchrony among rhythmical facial tremor, neocortical ‘alpha’ waves, and thalamic
non-sensory neuronal bursts in intact awake rats.Semba K, Szechtman H, Komisaruk BR.
A fine (approx. 9 c/sec) tremor of the jaw and/or vibrissae was observed in
normal rats while they were standing still and not showing gross bodily movement.
The tremor was distinctly different in frequency, intensity, and behavioral
context, from movements involved in gnawing, tooth chattering, or exploratory
sniffing. Individual tremor movements (recorded as EMG) occurred in synchrony
with individual bursts of multiunit activity (MUA) recorded in the ventrobasal
complex of the thalamus and with individual ‘spikes’ in the cortical
(frontal-occipital) EEG. Single trains of this rhythmical activity often lasted
more than a minute. The phase relationships between EMG and MUA differed among
individuals, but tended to remain consistent within each individual. Movement
artifacts were ruled out since (1) the moments of occurrence of individual tremor
movements and MUA bursts were interdigitated rather than simultaneous, and (2)
during high amplitude EMG bursts accompanying sniffing (associated with EEG
theta rhythm), tooth chattering, eating or licking, no corresponding activity in
the MUA was observed. We also ruled out the possibility that the neural activity
was generated by reafference, for (1) during vigorous non-tremor sniffing
movements of the vibrissae, or chattering or chewing movements of the jaws, the
rhythmical MUA was absent (although the units did discharge if the vibrissae
contacted an obstacle or were brushed by the experimenter), (2) rhythmical MUA
often continued both during brief pauses in the motor tremor, and in its absence,
and (3) injection of Xylocaine s.c. into the face abolished sensory responses of
the thalamic units, but the rhythmical MUA persisted. We discuss evidence which
suggests that (1) the rhythmical cortical EEG waves are the equivalent in the rat
of the alpha (mu) rhythm, and (2) the existence of parallels between alpha-tremor
and Parkinsonian tremor in terms of their mechanisms and functions. -
yes
-
