Talk:Texture from touch
This brief review of tactile texture perception is well written, but would benefit from attention to a few points, as detailed below. In general, the review tends to focus a bit too much on the work of two groups to the exclusion of other ideas. Condensing the thoughts currently presented would make room for discussion of the additional issues raised here.
It would help at the start to cite the superiority of touch over vision in texture perception (Heller, Percept Psychophys 45: 49-54, 1989) and the importance of textural information for haptic perception (Klatzky et al., J Exp Psychol Gen 116: 356-369, 1987).
I was surprised to see the thermal sense included as a dimension of texture. None of the papers cited in this context (Hollins et al., 1993; Hollins et al., 2000; Bensmaia and Hollins, 2005) actually endorse this view, and there is ample evidence for a distinct thermal sense, not only perceptually but also in terms of neural pathways.
There is no doubt that the series of elegant studies by the late Ken Johnson and his colleagues have clearly established spatial variation in the SA1 afferent population as a major coding mechanism for texture. The relevant studies are nicely reviewed by Bensmaia. However, the subsequent discussion of vibrotactile coding focusses exclusively on studies of Bensmaia and Hollins, ignoring other contributions that reinforce the idea of potentially multiple cues and multiple coding mechanisms. Examples are studies demonstrating the relevance of temporal cues (Cascio & Sathian, J Neurosci 21: 5289-5296, 2001; Gamzu & Ahissar, J Neurosci 21: 7416-7427, 2001) and tangential forces (Smith et al., Exp Brain Res 144: 211-223, 2002). There is also no attempt to integrate the ideas on vibrotactile coding with the concept of spatial variation as an exclusive code. It seems reasonable to admit the possibility that multiple cues and coding mechanisms are available, with the prominence of one or other perhaps depending on task demands, choice of exploratory strategy, and individual perceptual experience.
Finally, the discussion of cortical processing could include references to the lesion, neurophysiological and functional neuroimaging work of many others and go beyond S1, e.g., Randolph & Semmes (Brain Res 70: 55-70, 1974); Murray & Mishkin (Behav Brain Res 11: 67-83, 1984); Sinclair et al. (Somatosens Mot Res 13: 287-306, 1996); Jiang et al. (J Neurophysiol 77: 1656-1662, 1997); Roland et al. (PNAS 95: 3295-3300, 1998); Pruett et al. (J Neurophysiol 84: 780-797, 2000); Stilla & Sathian (Hum Brain Mapp 29: 1123-1138, 2008).
Some abbreviations are not defined (SA1, S1 etc).
---
This is a succinct, informative article on texture in primates.
The major issue that it overlooks is the fact that primates generally move their fingertip to generate signals. This is important and should at least be mention, if not fully discussed.
There are a few small issues. in the sentence "The main determinant of perceived roughness seems to be the spatial pattern of deformation of the skin (Taylor and Lederman, 1975) although temporal cues (Cascio and Sathian, 2001) and tangential forces (Smith et al., 2002) may also play a role.", Gamzu and Ahissar should be cited.
"The stimuli consisted of embossed dot patterns varying in their spatial properties. The roughness estimates, obtained for a variety of dot patterns, were plotted against predictions derived from each putative neural code." Please provide a few sentences about how the stimuli were delivered, both for the physiological and psychophysical testing.
In periphery, we have: "The spatial variability in the responses of slowly adapting type 1 afferents was found to account for perceived roughness of all the textures tested." while in cortex we have: "A subpopulation of neurons, whose SRFs or STRFs comprise excitatory and inhibitory sub-regions, effectively compute the spatial variability in the afferent input. In fact, a subset of these neurons exhibit responses to embossed dot patterns that match their perceived roughness: Their response increases for inter-element spacings up to about 2mm, then decreases (Arun Sripati, personal communication). " What are the current models for how these cortical responses arise from the peripheral ones? They seem to suggest very different coding mechanisms. Are they compatible?
In the parapraph on duplex theory of texture, where evidence of vibrotactile coding is first proposed, it would be an appropriate place to mention for the first time coding of texture in the vibrissal system. It is appropriate because in this system spatial coding is probably not critical (interwhisker distance varies from moment to moment, and rats can measure textures using a very small number) and the putative mechanisms are much closer to the vibratory model. For example, a statement like: "Indeed, the roughness of fine textures has been found to be highly correlated with the power of the vibrations these elicit in the skin, weighted by the spectral sensitivity of PC fibers." fits well with vibrissal coding, and the generalization is interesting. A quick link to http://www.scholarpedia.org/article/Vibrissal_texture_decoding would be helpful.
Reviewer A:
This revision has been quite responsive to the prior reviews. However, there are a couple of residual issues that merit attention.
The first two paragraphs still convey the impression that temperature is part of texture sensing. This should be clarified. Also, the value of the last paragraph on thermal sensing in an entry on texture is uncertain, and potentially misleading. I suggest deleting it. If, however, it remains in, it needs adequate integration and justification, and could profitably cite the correlative neurophysiological-psychophysical studies of Johnson and colleagues in the Journal of Neurophysiology (e.g. 1973; 36: 325 - 346 1973, 36: 347 – 370 1979; 42: 1297 – 1315 1979; 42: 1316 – 1331 1979; 42: 1332 – 1353).
Additional studies on potential non-spatial coding mechanisms are now cited, but only in passing, and no attempt has been made to integrate them into the discussion of complementary mechanisms, which focusses exclsuively on vibrotactile coding.
