Inversion effects in the expert classification of mammograms and faces

Abstract

A hallmark of a perceptual expert is the ability to detect and categorize stimuli in their domain of expertise after brief exposure. For example, expert radiologists can differentiate between “abnormal” and “normal” mammograms after a 250 ms exposure. It has been speculated that rapid detection depends on a global analysis referred to as holistic perception. Holistic processing in radiology seems similar to holistic perception in which a stimulus like a face is perceived as an integrated whole, not in terms of its individual features. Holistic processing is typically subject to inversion effects in which the inverted image is harder to process/recognize. Is radiological perception similarly subject to inversion effects? Eleven experienced radiologists (> 5 years of radiological experience) and ten resident radiologists (< 5 years of radiological experience) judged upright and inverted bilateral mammograms as “normal” or “abnormal”. For comparison, the same participants judged whether upright and inverted faces were “happy” or “neutral”. We obtained the expected inversion effect for faces. Expression discrimination was superior for upright faces. For mammograms, experienced radiologists exhibited a similar inversion effect, showing higher accuracy for upright than for inverted mammograms. Less experienced radiology residents performed more poorly than experienced radiologists and demonstrated no inversion effect with mammograms. These results suggest that the ability to discriminate normal from abnormal mammograms is a form of learned, holistic processing.

Analysis of Perceptual Expertise in Radiology – Current Knowledge and a New Perspective

“…Another popular model posits that initial global processing (consisting of bottom–up “global image statistics” like average orientation and average size of objects) signals if there is an abnormality (establishing its likelihood) without providing location information or constraining the subsequent serial search. The searcher can then change their strategy to a slower, more complete search for the abnormality (Evans et al, 2010, 2016; Drew et al, 2013a; Chin et al, 2018) (Figure 3).…”

Perceptual learning in the identification of lung cancer in chest radiographs

“…Because we do not have a precise measurement of the size of the tumor relative to the cutout, even the small cutout may have included some surrounding tissues, accounting for why performance was moderate but above chance. Although local contrast but not the larger background was important in our study, it remains possible that more extensive training may lead to greater reliance on the entire image (Chin et al, 2018). The forced-choice training procedure, with normal and abnormal images presented side by side, does not resemble radiological diagnosis.…”

“…As noted earlier, radiologists can rapidly extract the global image statistics to render an initial decision on whether a mammogram may be cancerous (Evans et al, 2016;Evans, Georgian-Smith, Tambouret, Birdwell, & Wolfe, 2013;Nodine et al, 1999). Additionally, presenting mammograms in an inverted orientation impairs radiologists' performance, suggesting that breast cancer detection is holistic (Chin, Evans, Wolfe, Bowen, & Tanaka, 2018). If perceptual learning of chest radiographs also involves the learning of global image statistics, then performance should decline when just the tumor cutout or just the background is presented.…”

Visual Search Performance and Strategy for Three-Dimensional Visualization Systems: Impact of Radiologist Experience

“…Particularly, radiologists are a limited resource [27], with a small amount of time available to dedicate to the experiment. For this reason, the number of qualified radiologists has been around ten individuals in most recent studies, e.g., [4,31,32,36,40]. We acknowledge that one cannot necessarily apply conclusions made in one study to the entire population.…”

“…Extensive training leads to the increased tuning of visual perception [28] and changes in search strategy which radiologists implement during image reading [29]. Most previous studies have demonstrated the strong top-down influence of expertise when comparing the performance of radiologists and non-radiologists viewing medical images [29][30][31][32][33]. In addition, the tracking of brain activity revealed significant differences among radiologists and non-radiologists, not only for radiology-related images, but also for control images [34].…”