Shifting and Asymmetrical Sharpening 1 The Attentional TemplateShifte - PDF document

1 Shifting and Asymmetrical Sharpening 1
Shifting and Asymmetrical Sharpening 1 The Attentional TemplateShifted and Asymmetrically Sharpened by Distractor ContextXinger Yu and Joy J. Geng University of California, Davis Author Note Xinger Yu, Department of Psychology and Centerfor Mind and Bra

2 in, University of California,Davis; Joy
in, University of California,Davis; Joy J. Geng, Department of Psychology and Center memory (round) in memory while looking for it. It is frequently assumed that visual search is most efficient when we hold veridical features that perfectly match the tar

3 get. However, this is not always templat
get. However, this is not always template) to bias sensory processing towards target largely assumed that the template is optimal when it perfectly matches the veridical target. However, linearly separable from the target(e.g., all were at the same orient

4 ation rotated counterclockwise from the
ation rotated counterclockwise from the target the target). Navalpakkahm and Itti (2007) found that the 60¡ stimulus was chosen more frequently than the 55¡ (true target) stimulus as the targetprobe trials. This demonstrated that the target distant Scolar

5 i and Serences (2009; Experiment 2, 3) u
i and Serences (2009; Experiment 2, 3) used an independent reasoned that contrast detection thresholds should be lowestfor orientation distant distractors are uniform and very distinct from the target. Similar findings of shifted target representations h

6 ave also irrelevant colored ÒcuesÓsurrou
ave also irrelevant colored ÒcuesÓsurrounding each possible target location. They found stronger attentional capture by red colored cues compared to orange ones, suggesting that attention was biased specific features (Becker, 2010; Becker et al.,Becker, F

7 olk, & Remington, 2010). Althoughthe pro
olk, & Remington, 2010). Althoughthe proposed mechanism differs from Navalpakkam and Itti (2007) and Scolari and Serences (2009), they also evidence hypothesized to decrease the selectivity of task Scolari & Serences, 2009). In our previous linearly separ

8 able). Two groups of subjects saw the sa
able). Two groups of subjects saw the same distractors (ranging from 5¼-60¼ along a color wheel), but the Òhigh-similarity groupÓ experienced a greater proportion of the most target Shifting and Asymmetrical Sharpening 8 high similarity group had a targe

9 t and was therefore very different from
t and was therefore very different from those used in other studies distractor values, but this will measure the contents of the target template. Trials from the two tasks wereinterleaved. The separation of the visual search training trialsand the templat

10 e probe trials for the presence of shift
e probe trials for the presence of shifting and asymmetrical sharpening in the target template due to the predictable than the target color) and could be predicted from trial-to-trial. could be from either the target students space (a,b was assigned a sin

11 gle target the target color. These distr
gle target the target color. These distractors werechosen to exceed the average just noticeable difference,yet be confusable withthe target when presented in a competitive counterbalanced across the distractors in the visual search training trials alwaysr

12 eferred to as positive rotations from th
eferred to as positive rotations from the target (i.e., 5¡, 10¡, 15¡) and the distractors that appeared only in the template probe trials (i.e. "untrained" colors) were Shifting and Asymmetrical Sharpening 11 In the bidirectional group, the distractors

13 in the visual search training trialswere
in the visual search training trialswere sampled from both directions of the target color: all analyses collapse across two target colors.both groups, the colors The target color was always present within Shifting and Asymmetrical Sharpening 12 In the co

14 lor template probe task (Figure probe tr
lor template probe task (Figure probe trials. Trials were Shifting and Asymmetrical Sharpening 13 processes beyond target representation, such as those necessary for resolving distractor competition. establish expectations for the target color and the pr

15 obe trials %&&&'('%)&&#*+ Shifting and A
obe trials %&&&'('%)&&#*+ Shifting and Asymmetrical Sharpening 14 of simultaneous distractor competition. Probe trials the estimated standard deviations are equivalent (see BCD=EBF (1) Because the response for each color in our probe task likelihood e

16 stimation), all parameters were estimate
stimation), all parameters were estimated using hierarchal Bayesian analysis (HBA). The hierarchical approach is particularly useful for this study given the small number of data points per subject because biasing the posterior distributions. towards the

17 template over ÒpositiveÓ colors. Recall
template over ÒpositiveÓ colors. Recall that negative colors were never seen directly compared the false alarm rates the Bonferonni method. Finally, in addition to null hypothesis testing, we also (BF)Rouder et al., 2009) for all student-t statistical a

18 nalyses using BayesFactor package in r (
nalyses using BayesFactor package in r (Morey, Rouder & Jamil, 2015). The BF is a statistical index of the evidence the data provides Shifting and Asymmetrical Sharpening 17 underlying template likely the µ value for the unidirectional group was more neg

19 ative group value was pos Shifting and
ative group value was pos Shifting and Asymmetrical Sharpening 18 Next, the ! values were there would be asymmetrical sharpening in the unidirection likely if the null hypothesis is true than the alternative. Together, these results are consistent with

20 asymmetrical next to the raw false alarm
asymmetrical next to the raw false alarm data as a more direct Shifting and Asymmetrical Sharpening 20 the bidirectional group had higher false distractors that are linearly separable from the target distractor sets from both sides of the target Shifting

21 and Asymmetrical Sharpening 21 Visual
and Asymmetrical Sharpening 21 Visual sets were used for a given participant in the unidirectional group whereas distractor sets were randomly interleaved trial-by-trial in the bidirectional group. Accuracy (Figure 5B) and reaction time (RT, the hypothe

22 se either the Nation the samethe unidire
se either the Nation the samethe unidirectional group in Experiment 1, withthe following exceptions: There were sequential blocks (Figure 6). The first (45¡, 40¡, 35¡) in block 3, (30¡, 25¡, 20¡) in block 4, and (15¡, 10¡, 5¡) in block 5. Distractor color

23 (negative or positive rotations from th
(negative or positive rotations from the target color) spurious were Shifting and Asymmetrical Sharpening 24 the colors that appeared only in the template probe trials (i.e. "untrained" colors) are labeled as ÒnegativeÓ values from the target(i.e., -5¡,

24 -10¡, -15¡). Procedure. The procedure
-10¡, -15¡). Procedure. The procedure of Experiment 2 was identical to Experiment 1, except for the additional factor of targetdistractor similarity, which increased in 5 blocks over the experiment (Figure 6). The order of the 5 blocks was held constan

25 t targetcolor. All error bars are SEM. S
t targetcolor. All error bars are SEM. Shifting and Asymmetrical Sharpening 26 tendency then entered into a one-way ANOVA with the factor with increases in distractor competition positive half of the target template was selectively sharpened over blocks

26 to counteract competition Shifting and A
to counteract competition Shifting and Asymmetrical Sharpening 28 Analysis of the false alarm rate. The raw false alarm rates for negative and positive non-targetcolors were enteredinto a 2 color direction (negative, positive) x 5 block (1-5) repeated me

27 asures ANOVA (Figure 8C). Recall that th
asures ANOVA (Figure 8C). Recall that the raw false alarmrates are not related to estimations of central tendency and are a direct likelihood of mistaking a non-target interaction was due to a linear decrease in positive false alarm rates over blocks, but

28 no change in negative false alarm wheth
no change in negative false alarm whether betterexclude highly similar distractors. Together, the results suggest that shifting and sharpening of the target this alternative hypothesis, Experiment 3 was identical to Experiment 2 (5 blocks of equal durati

29 on), but only the most dissimilar distra
on), but only the most dissimilar distractor set was used (identical to block 2 in Experiment 2). Method Participants. Twenty new UC Davis undergraduates (4 males, 1 left-handed, ages 18-27) participated in Experiment 2. Each Shifting and Asymmetrical

30 Sharpening 31 Design Experiment 2. The
Sharpening 31 Design Experiment 2. The central tendency of the target representation shifted away from visual targetcolor. All error bars are SEM. was not affected by practice over time. Figure Shifting and Asymmetrical Sharpening 34 Additionally, we

31 conducted a two-way ANOVA with analyzed
conducted a two-way ANOVA with analyzed !neg and !pos values as a function of block (Figure 12B). An ANOVA with simply to practice effe The results yielded a significant main effect of color direction, F(1, 38) = 5.36, p = .03, !p2 = .12, no main effect o

32 f experiment, F(1, 38) = .46, p = .50, !
f experiment, F(1, 38) = .46, p = .50, !p2 = .01, but a significant interaction, F(1, 38) = 13.52, p .001, !p2 = .26. Post hoc t-tests negative found in Experiment 2 was due to the change in the strength of competitionduring visual search and not practic

33 e. Analysis of the falselarm rate. Sim
e. Analysis of the falselarm rate. Similar to Experiment 2, the false alarmrates collapsed acrossdifferent negative and positive color degrees(Figure 12C) were entered (19) = 2.91, p = .045, d = .58, BF10 = 5.57),but none others (all Shifting and Asymme

34 trical Sharpening 38 early search trial
trical Sharpening 38 early search trials wereused to ÒtrainÓ the template. The template was then measured by a separate "template probe" task Shifting and Asymmetrical Sharpening 39 the targettarget. Using a separate probe task is necessary to measure t

35 he contents of the template, which are p
he contents of the template, which are presumed to be held in memory (Giesbrecht, Sy, & Guerin, 2012; Woodman, Carlisle, & Reinhart, 2013; Myers et al., ScolariSerences, 2009) in order to make inferences about how expectations for a visual search display

36 adjusts sensory taps into decisional pro
adjusts sensory taps into decisional processes as well, leaving open the possibility that our results reflect not only earlysensory changes, but also withinthe template (Wolfe, 2007; Smith & Ratcliff,2009). Importantly, the target is (in memory) from acti

37 ve selection during distractor competiti
ve selection during distractor competition This result may superficially appear to be at odds with the target and all three distractors were identical, possibly producing sharpening in the width of the targettemplate in responsethe strength of distractor

38 competition direct calculation of the fa
competition direct calculation of the false alarmrates to negative and positive non-target colors. In Experiment 2, we found that asymmetrical sharpening increased with distractor competition, suggesting that sharpening occurs to better templaare updated

39 in response to changes in distractor com
in response to changes in distractor competition Shifting and Asymmetrical Sharpening 41 degrees changes in the target pattern template on any given trial to select the target linearly separable distractor features, and an asymmetrical sharpening to incr

40 ease the precision of the target-to-dist
ease the precision of the target-to-distractor boundary (see suggeststhat template shifting increases signal-to-noise ratio by selectively sensory neurons tuned to distractor features (Reynolds & Heeger, 2009). Alternatively, the pattern we have seen may

41 alsoinvolve decisional processes involve
alsoinvolve decisional processes involved in determining the match between a stimulus and the attentional template held in memory (Smith & Ratcliff, 2009). Although address Shifting and Asymmetrical Sharpening 42 attentional biases in order to increase t

42 he representational distinctiveness of t
he representational distinctiveness of the target from expected distractors. In conclusion, our experiments reveal that the target multiple distractor features. Expectations regarding the linearseparability of the distractor set from the target produces

43 Shifting and Asymmetrical Sharpening
Shifting and Asymmetrical Sharpening 43 References Bauer, B., Jolicoeur P., Cowan, W. B. (1996). Visual search for colour targets that are or are linearly seperable from distractors. Vision Research, 36(10), 1439-1466. https://doi.org/10.1016/0042-6989

44 (95)00207-3 Becker, S. I. (2010). The
(95)00207-3 Becker, S. I. (2010). The Shifting and Asymmetrical Sharpening 45 Lee D. K., Itti L., Koch C., & Braun J. (1999). Attention activates winner-take-all competition among visual filters. Nature Neuroscience, 2(4), 375-381. http://doi.org/10.10

45 38/7286Ling S., Jehee J. F. M., & Pestil
38/7286Ling S., Jehee J. F. M., & Pestilli F. (2016). A review E., Rohenkohl, Shifting and Asymmetrical Sharpening 46 Scolari M., Byers A., & Serences J. T. (2012). Optimal deployment of attentional gain during fine discriminations. Journal of Neurosc (1

46 ), 223Ð231. http://doi.org/10.1016/j.neu
), 223Ð231. http://doi.org/10.1016/j.neuroimage.2008.07.043 Series P., Latham sharpening for orientation Automatic guidance of attention from working Shifting and Asymmetrical Sharpening 47 Stan Development Team.(2017). Package ÔRStanÕ. URL https://cran

47 .r project.org/web/packages/rstan/rstan.
.r project.org/web/packages/rstan/rstan.pdf &#x Tj ;T Q;&#x q 0;&#x.24 ;� 0 ;�.24;&#x 18 ;փ.;’ c;&#xm BT;&#x 50 ;� 0 ;P 1;І ;Ԡ ;&#xTm /;&#xTT1 ;
Tf;&#x 000;. Treue, 4.0: Current Progress with Oxford.Wolfe J. M. (20