User talk:Tobias Heed

From Scholarpedia
Jump to: navigation, search

    ROUND 3:

    I think you've made this a nice article now. The only thing that remains is that patient "DLE" is still named, though I'd say this is unneccessary.

    ROUND 2: I think that the article has vastly improved! It is now much more streamlined, and I have no major issues with the parts that were there before. I note below a few small things that could be edited to make final improvements. For the new section on external space, I have a few more comments, which I also list below. Comments go section by section.

    Hi Tobias, thanks for your comments and suggestions, our responses appear in-text below.

    Perceived location of touch plastic after cortical changes - this is correct, but wouldn’t you actually want to say “after cortical and peripheral changes”, as SI plasticity can be very strong not just after cortical changes such as stroke, but also after peripheral changes such as loss of a limb?

    We agree and have updated the text accordingly.

    Detection of touch and localization: A serial process. Sentence 1: I think a number is missing (there are major hypotheses - probably you want say “2 major hypotheses”

    Thanks for noticing this – we have changed it to “two major hypothesis”.

    criteria is plural, should be criterion (appears several times)

    We have changed most of these to criterion. “Criteria” remains for sentences that refer to both localization and decision tasks.

    In the sentence about “20mm”, “she” may be better formulated as “the patient in Paillards’s study” (as the reference is several sentences earlier); same in last sentence of that para (her numbsense)

    For clarity we have changed both of these references so that they refer to the Paillard study.

    Harris paragraph, 3rd to last line, issue is missing “the” last sentence of that para, “future cases… will need sample sizes” - probably you meant “future studies”, as a case would have N=1?

    Agreed, both of these have been updated.

    Localization in external space “what is called” could maybe be formulated a bit more elegantly, e.g. “usually referred to as” or sth like that?

    We have changed this to “usually referred to as the”.

    naming “DLE” seems unnecessary

    This has been removed.

    para Medina & Rapp, entence starting “They found”, “They” is an unclear referent same para, it remains unclear what the hand crossing result is supposed to tell the reader. I think that the many names of reference frames (eye-centered, body-centered, head- and body-centered) are still unclear at this point. It has not been described what “centered” means - that whatever stands before it is then the origin of the spatial coordinate system. In the Media & Rapp para specifically, it would be helpful to explain something along the lines of “if this effect were skin-centered, it should not change with different postures: a skin location remains the same, no matter how the respective body part is held”, and after reporting the crossing effect, “thus, the patient’s deficit was always stronger in one side of space, both relative to the head and the body, i.e., the effect was head/body-centered”.

    We have attempted to address these comments. A sentence is included to explain that “centered” means refers to the origin of a reference frames spatial co-ordinates. In regards to the Medina/Rapp para we have explained how the result speaks to the existence of external reference frames.

    para on behavioral studies”: First sentence is a bit weird. A paradigm is mentioned and then dropped (though I guess you did it this way to comply with my suggestion to include the link to the TOJ scholarpedia page - can this maybe go elsewhere?) Sentence starting with “Whereas”, this is only a half sentence and should be connected to the previous one with a comma (or start with “In contrast”)

    We have changed the topic sentence to reflect the content of the paragraph. The temporal order judgement page is now linked to as “see also evidence from temporal order judgement” (as we mostly have not used TOJ findings in this discussion).

    fingers in external space: I'm not so sure about this part. There is a lot of evidence that fingers, too, are localized with respect to external space. I think Azanon & Heed 2014 had a para on this question in it - there is some evidence for both (only anatomical, or both anatomical and external). The Haggard study has been addressed by a study by Riemer et al (I think in EBR) which called the Haggard study into question. Badde et al. 2014 have shown that fingers can be “assigned” to the wrong hand in a TOJ task when they are crossed over to that hand. Heed et al. 2012 have also shown evidence that finger localization uses external space (again TOJ). I'm not saying you should cite all my papers ;) My point would basically be that, although the fingers seem to be a bit weird, there is definitely evidence that they are also localized externally. You could even consider removing this part..

    We have adjusted this paragraph to emphasis that although the fingers seem different they are still localized in external (and not just somatotopic) reference frames (but perhaps not as influenced by posture as the hands according to Riemer).

    The Overvliet reference is definitely incorrect and should be replaced here. Overvliet 2011 is a tactile remapping paper and probably was supposed to go in the para one higher? In response to your response to my initial concern about that study: I had criticized stating that Groh and Sparks “examined” saccades to crossed hands. Overvliet DOES show such saccades; however, they are relatively few. Groh and Sparks had reported 10 curved saccades in 10 trials (i.e. 100%). Overvliet is nowhere near that. Besides this, these studies can of course be cited! The Overfliet results are well in line with Brandes 2015. Overvliet also gives some estimates for how long tactile remapping might take (though this would really clutter the para that already contains lots of numbers from Azanon and Brandes now…) - I hope this clears up the confusion...

    We have replaced the Overvliet reference with reference to the correct study. We have decided to not include the overvliet results in with the Brandes 2015 (I agree it would clutter the para).

    With respect to the “integration” hypothesis: in general for the “external space” section, maybe it could be made more clear in the beginning that probably the brain has all kinds of spatial info in different reference frames available at all times. The way you describe the Brandes study with the additional time used in the crossed condition reflecting integration is correct, but the point of Badde (and Heed) is that the brain always has several codes available, and weighs them to determine the location of touch. I understand if you don’t want to make the integration hypothesis a big point; but I think the point of “several reference frames in parallel” is a very general point that many researchers would subscribe to, and that could stand out more clearly in the beginning of the section.

    We have now linked to this point (in parallel processing) in the opening paragraph of this section. An additional sentence is included to emphasise that the Badde/Heed model includes weighting during integration.

    Open questions very last sentence, there is a scholarpedia article that could be linked,

    We have added in a link to the tactile illusions page.

    everything below here is from the previous round, and all responses and the respective changes to the article are ok for me

    ROUND 1:

    It is great that Scholarpedia will have an entry dealing with touch localization.

    Intro: “SI_somatotopic_maps" should probably be a link. it reads strange that first it is said that S1 maps exist, then the topic changes, then S1 is discussed again. Fig: is there a reason for the placement of the different boxes? why are some next to each other, others above/below? Fixed.

    in general, I can’t really see the red line here. Lots of things are mentioned (amputations, fingertips, serial vs. parallel…), but they don’t form a comprehensive whole. We have attempted to make these changes: “Given noise, both at the level of skin receptors and neurally, the brain needs to decide both whether a stimulus was presented and where that stimulus is located. … Finally, touch needs to be localized – not only to a position on the skin surface – but in reference frames relative to the body and the external environment. Therefore, the location of touch can be represented in a number of different representations, each with their own reference frame. We will present evidence on the different stages of processing and representation that are involved in localizing touch.”

    Detection… sentence 2, if this WERE the case para 2, sentence 2, “her” is an unclear reference with respect to the previous sentence grammar error in “not reporting feeling” Fixed.

    For a not so expert reader, it is difficult to understand here what is meant by “model”. The second “model” is only introduced as “participants first detect and then localize”. This is hardly a model (?), but rather a hypothesis. We’ve changed the text to reflect that it is a hypothesis (see also the new figure 2).

    Furthermore, given that tactile information contains localization information by nature (through identity of the activated sensors), and axons are very neatly ordered starting in the very first stations of anatomy in the spinal cord, a few words could be said about why one might then expect that there could still be a mere “detection” mechanism. I do not find this intuitive at all.

    We have included this in intro paragraph. We agree that location could arise from the orderly arrangement of the nervous system even though S1 is distorted. As long as S1 and the processes that read out activity from S1 are consistent. However such a model breaks down when S1 changes in response to use (Craig 1993) or cortical damage in stroke (which does not simply result in lost sense of touch).

    Figure text: the part about the experimenter is not understandable to a non-expert and should be more clearly explained end of section, 2 sentences with “now”, and it sounds as if participants modeled their own performance (“they" is an unclean reference) given the “dictionary” character of scholarpedia, it would seem adequate to mention signal detection theory and to link to a …pedia page that explains it. We have made these changes.

    The conclusion of the section is not clear to me. What the Harris experiment seems to say is that either I can detect AND localize, OR I can’t do either. Why does this imply seriality? To me, it suggests that the processes are not separated. The way things are explained is unclear. You say “detection & localization were equally sensitive”, which probably refers to signal detection theory? But this is unclear. It reads as if the two were now connected. What is meant by “detection was not dependent on localization”? Do you mean “participants sometimes detected stimuli although they could not localize them, but they never claimed to not have detected a stimulus and then were able to localize it”? This is a strong claim, given the binary nature of the responses (thus, I guess, the necessity to model the results to arrive at the stated claims)? >We have changed the text. Any reference to detection means conscious perception. We are describing conscious detection vs. the ability to unconsciously report on the properties of the touch (e.g. location). As for Harris, they provided three different hypotheses on how a stimulus can be detected and localized (see Figure 5 of Harris Thein and Clifford, 2004). In one hypothesis (A), the output of one “module” of sensory processing is both detection and localization information. This model would predict that noise would decrease BOTH detection and localization accuracy. A second model (model B in Figure 5) is that the output of a first stage of processing would result in information regarding detection, and a second stage output (receiving input from the first stage) would be used for localization. A third model (model C) would be akin to the “numbtouch” model, in which detection and localization are the result of parallel processes. The important difference between models A and B are the slope of the relationship between signal strength and d’ for detection and localization. For model A, the slopes are parallel; whereas for model B, the slope is higher for detection vs. localization (That is, for every unit of signal strength, detection accuracy improves more than localization in model B – whereas for model A, detection and localization accuracy improve at the same rate.) The observed results in Harris Thein & Clifford (2004) are more consistent with model B vs. models A and C (see Figure 4 A and B for an example).

    All that said, this is admittedly a bit difficult to distill in a Scholarpedia article! We’ve decided to avoid discussing the difference between model A and model B in this article – though we’ve referred them to the original article.

    Regarding the discussion section – this is referring to Experiment 3 and 4 of Harris Karlov and Clifford (2006). We have rewritten this to hopefully make it clearer.

    Plasiticity… sentence 2, these terms are not clear to a reader. “is” is grammatically wrong second para, could generally be written more accessibly for non-expert readers (asynchronous? pre-conditioning? controls?). >See text. Last sentence is uninformative. If you mention this, give us some clear examples. Otherwise it’s a lame claim. >Removed para 3, though I understand the “read-out” idea, I find it problematic, because it seems to imply the homunculus fallacy (a little man in the brain interprets brain activity…). Instead, further processing must take place. From this, conscious perception must somehow “emerge”.. Reference to future research is unnecessary. I find the last para clumsy. It could be written more clearly. We have changed this to “Given that S1 is plastic, the relationship between activity in a specific region of S1 and perceiving touch in a particular location on the skin surface cannot be fixed, such that one set of neurons always represents touch at a specific location. There must be further processing that takes information from somatosensory regions and interprets it, such that conscious perception of touch location emerges. Very little is known about exactly how the brain interprets somatosensory activity as a particular tactile sensation. Some initial evidence towards understanding this comes from individuals with brain damage due to stroke.”

    Body representations sentences 3 and 4 don’t go well together. “these” in sentence 4 has no clear reference in the previous sentence. para 2, beginning, what is meant by “supporting a representation”? “forward” is unclear, as when understood in a hand-centered reference frame it would imply that the hand appears to move into the head (which is how I understood it). Maybe say “away from the body”? “Or are the changes…” missing a “?” We have made these changes to improve the clarity of the text.

    ref to de Vignemont 2005: would a simpler explanation be that it could be a rational (rather than perceptional) conclusion that two stimuli are further apart when the finger is longer? This is a very interesting idea. I (Jared) think this is a definite possibility. For the purpose of this article, I do not feel comfortable stating that these results are “rational versus perceptional” without having some experimental evidence to support it. So for now, I would like to keep this section as is, while thinking about potential experiments that could address this question in my future research.

    lat sentence, why “location”? I have deleted location as the evidence only says that extent is changed.

    Localization… Fig 4: “Likely” does not start a new sentence, but a half sentence here (period before is incorrect) Fig 4: the claim is actually incorrect. Often, there is still a significant effect of crossing at very long SOA such as 1500 or 3000 ms. It is smaller (e.g. 100% correct uncrossed vs. 95% crossed), but consistently observed. We have removed figure 4 entirely.

    Fig. 4, the claim that “at 300ms” the order is “typically incorrect” is incorrect. There is huge variance in this: a) whether participants actually show a systematic reversal (i.e. are MORE incorrect than correct: many participants are a bit less correct, but still MORE correct than incorrect), b) the time of a potential reversal is variable.

    We agree that there is definite variance in participants responses, with some showing the “flip” (e.g. Y&K 2001, Figures 1 and b) whereas others do not (Figures 1 c and d). We also agree that the reversal varies according to subjects (e.g. Figure 1a versus Figure 1b). This has been changed.

    the reference to Groh and Sparks is unfortunate. This paper tested totally different things. It’s a monkey study. As one idea, the authors then tested themselves (!! not the monkeys) on something like 10 trials (paper states “5-10 minutes” and reports 10 saccades each for uncrossed and crossed in Fig. 6) with crossed hands and report this in the paper. I know this ref is cited frequently, but it’s just wrong to say that the authors “examined” this. They reported a very preliminary finding which was qualified in many ways later. For instance, the 10 reported saccades in the crossed condition all seem to curve, and this has not been replicated (see Overvliet paper).

    We understand the concern. We should have also discussed the Overvliet paper. We think that the curved results were seen in the Overvliet paper (e.g. see Figure 1, immediate response, crossed and Figure 3). Perhaps I misunderstood – are you discussing the type of curves? However, to remove any potential confusion, we have removed this discussion from the passage.

    The logic of why the TOJ would say something about localization is not really clear in the description. Finally, the paragraph represents one (outdated) of several views of tactile remapping. The reference list contains Badde & Heed 2016, but our ideas and experiments are not discussed. I find this disappointing. We have presented lots of evidence in the last years that imply that tactile localization relies on the integration of spatial information coded in several spatial reference frames that are available in parallel, and that remapping is NOT (only) a serial process as has been suggested by Kitazawa and is implied in this para here. We have provided numerous findings that do not rely on the TOJ (e.g. Brandes & Heed 2015 JNeurosci; Badde et al. 2016 Psychon Bull, see summary in the 2016 Badde & Heed review). Please note that especially the “300ms” issue stems entirely from the 2001 Kitazawa study and is related solely to TOJ. For a brief description of the different accounts of remapping, see e.g. second para of intro in Brandes & Heed 2015, for a more extensive discussion Badde & Heed 2016 p7ff., and/or Heed & Azanon 2014 p5. And last, I recall seeing an entry from Kitazawa on tactile temporal order – it might be nice to link to it from here (despite my argument that TOJ are not the whole story)

    We feel that these concepts are too difficult to get across in a Scholarpedia article. Instead we have focused on showing that there is good evidence for the existence of different reference frames in within which touch can be represented. We link to the tactile temporal order section and also reference different mechanisms of how touch could be localized into an external reference frame. We have included the new integration account. A comparison of serial remapping accounts and the integration account would be a good fit for the Scholarpedia page on temporal order judgment.

    Open questions seem very detailed (numbsense) - such ideas could be moved to the respective sections. Additionally, para 1 reads like “oh by the way, everything we said at the beginning is actually maybe not true”. This, too, should be moved to the respective section. For open questions, it would seem more adequate to focus on broader questions. Examples may be (admittedly, this is very subjective): are “representations” a good way to conceptualize perceptual (body) processes? Are box and arrow models a good way to describe brain function? if so, which kinds of representations do we have to assume? But, can we come up with other concepts that better explain the phenomena we observe? Also, I do like that you pick up the “short vs. long lasting changes” - for me, one puzzle is actually the opposite than what you describe: that the brain is willing to perceive very strange things like 30cm long noses (which we NEVER have experienced as reality) after only few seconds of sensory stimulation, i.e. the great flexibility the brain allows itself to represent its body. I find the outlook to try and help patients a very good point for the open questions (maybe extend a bit?) Btw, right now, the open questions go very far into body representation, though the title of the article is “location of touch”. We have tried to address this in the section, although it is difficult to know where future research efforts should focus. Title: User talk:Tobias Heed Page ID: 42639

    Personal tools

    Focal areas