Abstract
A target letter is usually more difficult to identify when flanked by other letters. This phenomenon, known as crowding, strongly depends on the similarity of the target and the flankers: reducing target-flanker similarity normally benefits target identification (‘uncrowding’). For example, when reporting the central letter of a trigram, performance is superior when the trigram alternates in contrast polarity (e.g., black, white, black) compared to a trigram with uniform letters (e.g., white, white, white). Here, we show that alternating contrast polarity is detrimental when reporting all letters of a trigram. Stimuli consisted of uniform and alternating trigrams, presented for 150 ms at 10 degrees eccentricity. Participants reported either a single letter or all three letters. In the single report, performance on the central letter was superior in the alternating compared to the uniform condition, replicating the well-known advantage of ‘uncrowding’ by opposite contrast polarity. On average, the accuracy of all three letters did not differ between alternating and uniform trigrams in the single report. By contrast, when reporting all letters, average performance was superior in the uniform compared to the alternating condition, reversing the usual pattern of results when only reporting the central target. There was no advantage for the central letter in alternating compared to uniform trigrams when all letters were reported. Our results show that the benefit of low target-flanker similarity observed in standard crowding paradigms can be reversed by changing the task requirements. We suggest that benefits of uniformity outweigh the deleterious effect of crowding when all letters are targets.
Acknowledgement: Swiss National Science Foundation (SNF grant PP00P1_163723).