- Jul 2018
-
europepmc.org europepmc.org
-
On 2016 Apr 08, Lydia Maniatis commented:
This article is a follow-up to Gheiratmand, Meese & Mullen (2013) and is vulnerable to the same criticisms, which can be found here: https://pubpeer.com/publications/23283693
Other comments: In the first sentence of the article the authors state that: "The processing shape and form begins with the encoding of local orientation information in the visual scene by arrays of neural mechanisms selective for different orientations."
What is the basis for this statement? The local orientation of what? Of points? Of horizontal rows of points, vertical rows of points, diagonal rows of points, points having identical luminance, points varying in their luminance? How many points per locally oriented item? On what principle are the points linked into rows, so that they can be "detected"? How about subjective contours, or curves? How about orientations in 3D? How about amodal contours?
Given the well-established fact that non-local, high level principles mediate the percept, the notion that the percept can reveal "low-level" detector mechanisms and their "tuning" lacks, as Teller (1984) would put it, face validity. Even threshold effects have been clearly shown to be stimulus-dependent, i.e. the results of "tried and true" Gabor patches don't generalise.
In other words, the authors foundational claim is as obsolete and invalid as it could possibly be. But they assert it, and proceed accordingly.
The absence of a defensible rationale is (as is typical in this category of studies) complemented by a casual approach to assumptions in general (caps mine). Thus, "a von Mises function is used in part because "it IS THOUGHT TO BE the best function to fit to neurophysiological orientation tuning data. (Swindale, 1998)." So basically, Swindale thought this 1998, and Gheiratmand and Mullen apparently think it because Swindale thought it. Nuff said. Similarly: "m is fixed at 4, which has been CONSISTENTLY USED in the literature for approximation of the probability summation rule." I'm sure those other people had a good reason to use it. And: "we use a model involving arrays of orientation and spatial frequency tuned filters whose outputs are combined across the visual extent of the stimulus using a Minkowski summation rule. This method HAS BEEN USED previously to determine orientation and spatial frequency tuning for achromatic stimuli."Well, as long as its been used before. Some people must have thought it was good enough. Summing up: "We have used the classical psychophysical method of subthreshold summation to measure orientation tuning of the visual detectors underlying human colour vision at different spatial frequencies." Because old (classic) is best, e.g. some people think classical music is better than modern, or that old Coke tastes better than new. And if our: "visual detectors at different spatial frequencies" assumption is invalid (which it certainly is), our experiments will never reveal it, thanks to our other insulating, arbitrary (though confirmed by popularity) assumptions and restricted choice of stimulus features, that allow us to interpret our results in terms of those features (whether they are, in fact, relevant or not).
This comment, imported by Hypothesis from PubMed Commons, is licensed under CC BY.
-
- Feb 2018
-
europepmc.org europepmc.org
-
On 2016 Apr 08, Lydia Maniatis commented:
This article is a follow-up to Gheiratmand, Meese & Mullen (2013) and is vulnerable to the same criticisms, which can be found here: https://pubpeer.com/publications/23283693
Other comments: In the first sentence of the article the authors state that: "The processing shape and form begins with the encoding of local orientation information in the visual scene by arrays of neural mechanisms selective for different orientations."
What is the basis for this statement? The local orientation of what? Of points? Of horizontal rows of points, vertical rows of points, diagonal rows of points, points having identical luminance, points varying in their luminance? How many points per locally oriented item? On what principle are the points linked into rows, so that they can be "detected"? How about subjective contours, or curves? How about orientations in 3D? How about amodal contours?
Given the well-established fact that non-local, high level principles mediate the percept, the notion that the percept can reveal "low-level" detector mechanisms and their "tuning" lacks, as Teller (1984) would put it, face validity. Even threshold effects have been clearly shown to be stimulus-dependent, i.e. the results of "tried and true" Gabor patches don't generalise.
In other words, the authors foundational claim is as obsolete and invalid as it could possibly be. But they assert it, and proceed accordingly.
The absence of a defensible rationale is (as is typical in this category of studies) complemented by a casual approach to assumptions in general (caps mine). Thus, "a von Mises function is used in part because "it IS THOUGHT TO BE the best function to fit to neurophysiological orientation tuning data. (Swindale, 1998)." So basically, Swindale thought this 1998, and Gheiratmand and Mullen apparently think it because Swindale thought it. Nuff said. Similarly: "m is fixed at 4, which has been CONSISTENTLY USED in the literature for approximation of the probability summation rule." I'm sure those other people had a good reason to use it. And: "we use a model involving arrays of orientation and spatial frequency tuned filters whose outputs are combined across the visual extent of the stimulus using a Minkowski summation rule. This method HAS BEEN USED previously to determine orientation and spatial frequency tuning for achromatic stimuli."Well, as long as its been used before. Some people must have thought it was good enough. Summing up: "We have used the classical psychophysical method of subthreshold summation to measure orientation tuning of the visual detectors underlying human colour vision at different spatial frequencies." Because old (classic) is best, e.g. some people think classical music is better than modern, or that old Coke tastes better than new. And if our: "visual detectors at different spatial frequencies" assumption is invalid (which it certainly is), our experiments will never reveal it, thanks to our other insulating, arbitrary (though confirmed by popularity) assumptions and restricted choice of stimulus features, that allow us to interpret our results in terms of those features (whether they are, in fact, relevant or not).
This comment, imported by Hypothesis from PubMed Commons, is licensed under CC BY.
-