The patient experience of diagnostic error

Graber, Mark Mark.Graber at VA.GOV
Wed Oct 30 13:17:47 UTC 2013

David – thanks for a very interesting explanation of the illusion, but I was most intrigued by your comment that inferential processing was quite different.  And I’m sure it is, but it seems to me that the illusion raises two points that intersect with inferential processing, somehow:

 1.  The importance of context – The visual illusion seems to be highly dependent on context.  The context-dependence is a key feature in inferential reasoning as well, and context errors are so famous in aviation and medical disasters, as you well know and have written about
 2.  In both perceptual and inferential errors, there seems to be a common pathway through the ‘feeling of right’ center where the “HOLD IT!” signal doesn’t arise to prevent the error for some reason.  I just have the sense that if we had a better understanding of this cognitive inspection step, and could improve it somehow, that it would be harder to fool us with visual illusions and easier to discover the instances of faulty reasoning that arise from System 1 and currently get the ‘ok’ sticker from the ‘feeling of right’ center.  This defective inspection may be one of the most common causes of cognitive diagnostic error.


Mark L Graber, MD FACP
Founder and President, Society to Improve Diagnosis in Medicine
Senior Fellow, RTI International
Professor Emeritus, SUNY Stony Brook School of Medicine

On 10/27/13 10:21 PM, "David Woods" <woods.2 at OSU.EDU> wrote:

The post below refers to a classic illusion, but this along with many others, are not examples of how the human mind can be confused. Quite the contrary, this example reveals important and powerful aspects of how the human perception functions in a complex, ambiguous, and dynamic world.

The illusion in question illustrates that there are limits to the remarkable property of the human perceptual system called lightness constancy (and color constancy).  Lightness and color constancy are a examples of competencies of human perception that computer vision algorithms struggle to produce, and that perceptual scientists only understand partially (yes, there are limits to human perceptual capabilities, as there are for all physical systems).

The human perceptual system uses relationships about surface reflectance instead of surface luminance. Changes in illumination therefore do not change what we perceive as a whole and stable surface. This is in stark contrast to the current state of computer vision and machine vision models and algorithms. These typically operate on luminance and hence interpret changes in illumination as changes in surface. For example, if the illumination source were to move in this illusion, the movement of the shadow would be detected as change in the surface by many if not all of existing computational vision models. In order to manage this problem, these models will assume properties of the surface or of the illumination source, or the stability of the environment over time. For instance, one method would be to collect images of changing illumination conditions and then use these changes to compute a reflectance map, which can then be used to assess and remove future variation in illumination. Of course, this technique requires changes in illumination to determine what is constant (reflectance of the surface present over time in the scene) and what varies (luminance varies with illumination and viewing angle). The mind quickly extracts patterns in reflectance that allow direct apprehension of properties of the scene despite huge changes in luminance as sources of illumination and viewing angle change -- a remarkable capability (as a result, few people even appreciate the extreme difficulty of the problem).

To clarify, the illusion cited helps us notice what under realistic conditions is a special human perceptual competency, and second the illusion does not address inferential processes.  How inferential processes work, and break down, is quite different.

David Woods, PhD  木材
Releasing the Adaptive Power of Human Systems

* Lead, Initiative on Complexity in Natural, Social & Engineered Systems
* Co-Director, C/S/E/L Cognitive Systems Engineering Laboratory



To unsubscribe from IMPROVEDX: click the following link:

Visit the searchable archives or adjust your subscription at:

Moderator: Lorri Zipperer Lorri at, Communication co-chair, Society for Improving Diagnosis in Medicine

To learn more about SIDM visit:

To unsubscribe from the IMPROVEDX:
or click the following link: IMPROVEDX at LIST.IMPROVEDIAGNOSIS.ORG


For additional information and subscription commands, visit:

http://LIST.IMPROVEDIAGNOSIS.ORG/ (with your password)

Visit the searchable archives or adjust your subscription at:

Moderator: Lorri Zipperer Lorri at, Communication co-chair, Society for Improving Diagnosis in Medicine

To unsubscribe from the IMPROVEDX list, click the following link:<br>
<a href="" target="_blank"></a>

More information about the Test mailing list