Talk:Efficient coding hypothesis

This topic is being edited as an assignment in an undergraduate neurobiology course. The course is participating in the Wikipedia Education Program. The revised article will be posted by March 24, 2014.Iutschig (talk) 22:54, 18 February 2014 (UTC)

Note for the reviewers: The efficient coding hypothesis had a few review articles, but it is best explained by actual experimental results and procedures. Therefore, we included primary sources in order to give readers an understanding of how coding of natural image statistics has actually been observed in real neurons. Iutschig (talk) 14:57, 27 March 2014 (UTC)

Secondary Review: Overall, I think you did a fantastic job explaining the efficient coding hypothesis as well as the criticisms of the hypothesis. These two sections were my favorite parts because you methodologically explained what and why. I do have a few recommendations to help improve your article. First, I would suggest adding in a few pictures. Looking on the WikiCommons is a great place to get started and you don’t have to worry about copyright issues. I think a picture of a basic neuron and its structure would be useful for your article. A picture of an action potential would also be interesting to look at, plus it would break up some of the text heavy areas. Second, I saw a few references (in the evolution section and the natural images and statistics section) where the author’s names were placed in parentheses after the sentence. I would recommend either using the researcher’s names in the sentence or using the “add a source” icon to put in a superscript. One of my favorite parts of the article was the last section title biomedical applications. Is there information about how changes in the design of the cochlear implant could increase speech intelligibility? I think it is awesome that it can do that, but I am curious to learn more about how! I think that could be interesting for readers (if there is even information available about that topic). Overall, I think this article sounds very good! I am just offering a few suggestions about minor parts of your article. Great work! KPhillips13 (talk) 02:24, 28 March 2014 (UTC)

Primary Review
This article is very well-written and the information flows nicely. The sentence structure and wording is clear and concise. Spelling and grammar are almost perfect. However, I did notice that in the section “Hypothesis for Testing the Efficient Coding Hypothesis,” the last sentence reads, “researchers should observe is that there…” In this case, the ‘is’ should not be included. Also, in the section “Examples of these Approaches” under 2, the fourth sentence is missing the word ‘of’ when it say, “… sparse coding during the viewing these natural scenes.” Also, in the twelfth sentence of this same section, the phrase ‘more greatly’ is used, which does not sound like proper grammar. The last grammar mistake I found was in the section “Criticisms of the Efficient Coding Hypothesis” under “observed redundancy.” The second sentence reads, “Simoncelli notes that one major argument of critics in that higher up…” I think that you mistakenly wrote ‘in’ instead of the word ‘is.’ As for the compliance to the manual of style guidelines, you did a great job. The lead section provides a great background and overview about the efficient coding hypothesis. You specifically followed the guidelines by bolding the article title in the first sentence. However, one suggestion would be to link more terms in the lead section. For example, if possible, I would try to link words such as neuroscience, electrical impulse, neural code, sensory system, and visual/audio information. This would provide readers with additional background information that relates to the topic. Also, the layout of this article is organized and the section headings provide a good introduction to the material that is subsequently discussed. However, similar to the suggestion for the lead section, I would suggest linking more terms. Any important term or phrase that could potentially be unknown to the reader should be linked. While you do not want to link too much, you do want to provide the readers with information that will enhance their understanding of the topic. As stated in the manual of style, there should not be unsupported attributions such as “many scholars say,” or “it is said.” You did a great job of attributing the ideas in the article to specific authors. However, there were a few instances when you wrote, “Researchers have found..” I would suggest possibly attributing these findings to the specific researchers. Overall, however, the article is very well-written.

The article is also good because it is verifiable, however it does contain a minimal amount of original research. All of the information presented in the article appears to come from reliable sources, which can be accessed by the reader. There did not appear to be any instances where facts/ideas were presented without support from a published source. All of the article’s references are properly formatted in the “Reference” section. Also, you did a good job of including in-line citations for all the information you obtained from the given sources. However, there were a few time when the authors names were in parentheses at the end of a sentence. For example, in the section “Evolution Based Neural System,” you have (Ming & Holt) at the end of a sentence. This also occurs in the sections “Natural images and statistics” and “Hypothesis for testing the efficient coding hypothesis.” Wikipedia does allow the use of parenthetical citations, however, they suggest only using one type of format. Therefore, because most of the article already uses the hyperlinked, superscript box format, I would suggest deleting those parenthetical citations. Also, you added an inline citation to the section heading “Criticisms of the Efficient Coding Hypothesis,” which I do not think is necessary. In all other instances, however, you did use the proper Wikipedia format for in-text citations. You mention that you used primary sources in order to give the readers a better understanding of the topic. Overall, I think that you did a good job of incorporating secondary sources. Although Wikipedia prefers the use of secondary sources, they do allow you to use primary sources. In compliance with Wikipedia, it seems like you made good use of the primary sources. For instance, the article heavily relies on your secondary sources, but the primary sources are used backup this information. Therefore, for your article, I think the use of the primary, original research is acceptable.

In terms of the information covered in the article, I think you did phenomenal work. It appears that all relevant material is addressed and thoroughly discussed. One suggestion is to expand the “Biomedical Applications” section. If it is possible to find more information about this particular aspect, I think it would really enhance the article. Overall, there did not seem to be any major aspects about the topic that were excluded or improperly presented. I really liked how you described the methodological approaches and then gave specific examples of it. You mentioned on your talk page that you used primary sources to illustrate these approaches. Although Wikipedia advises you to be careful about primary sources, I think you made good use of them. In addition, I liked how you presented the criticisms for the efficient coding hypothesis. I think you wrote a great, well-rounded article which touches on all the necessary aspects of the topic. Moreover, I think that the article is written from a neutral viewpoint. I think you achieved a neutral point of view by including the criticisms of the efficient coding hypothesis. In this way, the reader is presented with all views regarding efficient coding hypothesis and they are not swayed into thinking a certain way about the topic. Also, by presenting the criticisms, it shows that the article is not written from a biased perspective. The use of the specific research studies in the “Examples of these approaches” section are potentially bias, however, I think you balance these sources. The criticisms help to balance any potential bias of the primary research studies. Overall, I think you did a great job of presenting a neutral point of view in this article.

The article would be significantly enhanced with the use of images. Providing images will help the reader understand the topic and also increase the aesthetics of the page. One suggestion is to include an image of the visual sensory pathways. You could show an image of an individual neuron or retinal ganglion cell which are discussed throughout the article. Another option would be to simply show how neurons communicate and transmit signals. This would give the reader a better understanding of the actual mechanisms of the sensory systems. Once you find an appropriate, related image, just make sure to properly cite it according to copyright laws and give an informative caption.

One of your sources entitled “Efficient coding of natural images” by MA Li-Bo and WU Si is a great secondary review article. It is properly cited throughout the article, especially in the section “Natural Images and Statistics.” You did a great job of including the review article’s information about statistical models of linear and nonlinear features. However, after reading the review article, I think that some of the information you presented in your article could be expanded. For example, the review article discussed models such as TICA and MAP for nonlinear features that you did not mention in your article. Also, I think you could mention more details about the hierarchical variance model that was conducted by Karklin and Lewicki. Also, I think you could include more information about the color and time methods which are discussed in the fourth section of the review article. Overall, I think you did a great job of incorporating this review article into your Wikipedia article. However, I think if you just expand on the main statistical models a bit more, then the reader will gain a better understanding of how they relate to neurons in the visual system.Smorrissey7 (talk) 20:55, 29 March 2014 (UTC)

Response to this Primary Review
Thank you so much for your in-depth review! The grammar notes were very helpful-we are correcting these. As far as adding links, we are planning to go throughout the whole article and add links for anything that requires further explanation. We are also planning to expand the cochlear implants section: the source referenced in that section does have more information to include about cochlear implant design, and we are going to look into creating a summary of how efficient coding algorithms can be applied to cochlear implant design. Also, thank you for suggesting particular images that we could use: we were not sure what images would actually be helpful before your suggestions so we will look into finding images such as this in the Wikimedia Commons. You mentioned expanding on TICA and MAP models for nonlinear features, as well as including more information about the color and time methods. We are going to add some more information on these subjects, but will probably rely on linking since we think that actually understanding the algorithms for coding is beyond the scope of our understanding at this time. We are glad that the use of research studies was helpful: we thought that this helps to make the hypothesis more concrete. Thank you for your feedback.Iutschig (talk) 20:36, 10 April 2014 (UTC)

Another Primary Review
The article was well written. I crosschecked what was referenced from the fourth source, authured by Li, with the source itself. The overall points in the article were correct, but I noticed that there was some errors. The page states that the optic nerve transmits information at 10^6 bps, but the source sites 10^7. It might also be worth noting that the further reduction down to 40 bps is the cause of inattentional blindness, in addition to when it is mentioned later in the section. It may also be worth using some more of the information from this source in the Visual Systems section. For example, it is worth noting that visual processing includes what the eye is focusing on in addition to responding to events.

The article hits the criteria for a well written article. The article is clear and concise, and the sections are all written to contain a good level of explanation. The sources are not all secondary sources, but the reason for that was given and makes sense. The topic seems to be covered in a good amount of depth and breadth, covering multiple approaches to the hypothesis, criticisms to it, and applications of the hypothesis. Nothing stuck out as being biased toward one particular viewpoint; everything seemed to be neutral.

There are two additions that would help the article. The first is to add some images to help show explain some of the points, as was mentioned in the other primary review. For example, there are a number of images in the Li source that demonstrate the loss of definition in an image which may help demonstrate the point laid out in the Constraints of the Visual System section. The other addition is to add some more blue links. The article only contains blue links to other pages in the introduction and the first section. After, there are none. The article would benefit from the addition or spacing out of blue links throughout the whole article. Zxdsqw (talk) 20:52, 4 April 2014 (UTC)

Response to Primary Review
Thank you for your feedback! First of all, we double-checked the statistic we gave about the speed with which the optic nerve processes information. It is indeed 10^6 bps. Additionally, we clarified that the further reduction to 40 bps results in inattentional blindness farther up in the “Constraints on the Visual System” section in order to link that information together. Additionally, we will add more information from the Li article to this section in order to expand it further as this is a foundational section for understanding the efficient coding hypothesis. While we would love to include images from the article itself, it would be difficult to get approval to put them on Wikipedia since they are not our intellectual property. Consent would need to be given by the authors and this would take longer than the course allows. However, we are planning on putting in a few images from the Wikimedia commons in order to better illustrate the topic. Finally, we agree that the links should be throughout the article and we will add them in. Once again thank you for your constructive comments! EmmaAWeber (talk) 20:53, 10 April 2014 (UTC)

Secondary Review
This article is very well written and flows very smoothly- I think you guys did a really commendable job on what you put together. You made this topic "easy" to understand when reading it and I think that's an important part of the project. I only have a few suggestions- you guys nailed it mostly so here's what I can think of for the moment- add an illustration. Its easy to do and would enhance the page. Even a visual of the guy Claude Shannon- sometimes its cool to see the face of the guy/girl who discovered the whole thing. You could of course go for a more scientific figure- but anything would be good. I would also suggest that you change the title that says "examples of these approaches"- in fact you could just take that out and go on to 1. right from the end of the sentence saying ...statistical properties to this response". It just gives the article a more formal tone in my opinion. I liked the way there were a lot of links and citing going on. Maybe expand on biomedical applications? I feel like there are probably more things you could say about it vs. just one sentence. Other than these things you guys did a fantastic job!

Mady mads (talk) 04:54, 7 April 2014 (UTC)

Secondary Review
This article was very interesting and overall I really enjoyed reading it. However, there are a couple things that could help improve the article. There were several parts in the article that could have really used an image to help support and visualize what was being addressed. An image, along with a caption, could have made explaining an action potential and testing hypotheses easier to understand, especially for a reader who would have a limited background in neurobiology. The way the article is divided into subsections showed progression and presented the topic in a very organized way. It made the article very easy to follow. I specifically enjoyed the “Criticisms of the Efficient Coding Hypothesis” section because it presented this article without bias by showing all aspects of this method. --Mchan19 (talk) 02:02, 8 April 2014 (UTC)

Secondary Review
I think that your article had a really good flow and was set up perfect. I thought it was very organized and made it easy for the reader to understand what was being said. As stated in a few of the other reviews, I agree and think you need to add some images to aid readers that do not have a biology background. I thought the "Examples of these Approaches" section was very well written but think if you added images it would be even better! Over all you have a well written and well researched page! --Jordannetts (talk) 04:08, 8 April 2014 (UTC)

Secondary Review
I enjoyed reading this article. I agree with the other reviewers in that a few images could help to explain the topic and show a face or two. The images would help with those who don't understand much of the technical aspects of the topic and help the flow of the page because it may seem a bit dense to read about all at once. Overall though you did wonderfully and made great use of the primary sources to help explain the actual observations made. Johnsep12 (talk) 04:53, 8 April 2014 (UTC)

Secondary Review
Well done! This article is not only well-written, but it provides a substantial amount of important information. Even though we aren't really supposed to use primary sources, I think you all used your sources in a way that kept the article neutral. Maybe it is just because I appreciate reading about experiments, but I really enjoyed your "Examples of these Approaches" section. You all thoroughly explained all there is to know about this topic. My only criticism would be to shorten the titles of your sections. Overall, you all edited this article very effectively, in my opinion. -- MUbrooke31 — Preceding undated comment added 05:16, 8 April 2014 (UTC)

Another Primary review
Overall this was an interesting article to read. It was very comprehensible and there were very few grammar mistakes. there was a good stream of idea that explained the subject in an appropriate manner. I only have a couple suggestions about formatting. In the section titled, "Evolution based neural systems"it would be easier to read if the three major concepts were indented and listed rather than reading them as a long sentence. Another suggestion would be to make the section, "Examples of these approaches a subsection of "Methodical approaches for testing the hypothesis." Also in the example section, is it possible to link: circular receptive fields, inferotemporal neurons, and multi electrode array? Other than those minor changes, the article is fantastic to read!

The article did not get too specific or technical, which is a good thing. The amount of examples used was large enough that the readers could understand the practical purpose of the hypothesis, but the article stayed focused and stayed on point. The article was also very neutral. The didn't seem to be any implied bias from the language used, and there was a very formal tone used. There aren't many visual aids and i can understand why, but it would nice to see what kind of equation they are using to test this hypothesis, there could also be flow diagrams about how the images are processed, and maybe a picture of the circular receptive fields in the retina.

The article I used to check references is called "Vision and the statistics of the visual environment" by Eero P Simoncelli. This is a good article that explains the use of the Efficient coding hypothesis very clearly. It is a good article because it is written very well and doesn’t jump to conclusions quickly. He points out that the efficient coding hypothesis is by no mens a complete theorem and there are still many improvements to be made to properly discern the neural code for sending and processing information. The Wiki article did a great job of quoting the article and there didn’t seem to be any misrepresentation of information in the article. Overall great job guys i enjoyed reading it.7243HODGSOZ (talk) 20:57, 9 April 2014 (UTC)

Response to Primary Review
Thank for writing the primary review about our article. Your feedback was definitely appreciated. We are glad that the article was not too technical and easy to understand. We also thought that our examples would be a good fit for this article. We agree about the "Evolution based neural systems" section and that the three concepts should be made into separate lines. We will also merge the "Examples of these Approaches" into the "Methodological Approaches for Testing the Hypotheses" section in the heading. We did not get a chance to put links in or pictures so we will also do that in the article. We will go through the entire article and put in links where they are necessary and need to be explained. Thank you for taking your time and writing about the article. Muhippolover (talk) 20:22, 10 April 2014 (UTC)