Conceived and designed the experiments: MWD DB. Performed the experiments: MWD PCH. Analyzed the data: MWD. Wrote the paper: MWD PCH DB.
The authors have declared that no competing interests exist.
Early deafness leads to enhanced attention in the visual periphery. Yet, whether this enhancement confers advantages in everyday life remains unknown, as deaf individuals have been shown to be more distracted by irrelevant information in the periphery than their hearing peers. Here, we show that, in a complex attentional task, a performance advantage results for deaf individuals.
We employed the Useful Field of View (UFOV) which requires central target identification concurrent with peripheral target localization in the presence of distractors – a divided, selective attention task. First, the comparison of deaf and hearing adults with or without sign language skills establishes that deafness and not sign language use drives UFOV enhancement. Second, UFOV performance was enhanced in deaf children, but only after 11 years of age.
This work demonstrates that, following early auditory deprivation, visual attention resources toward the periphery slowly get augmented to eventually result in a clear behavioral advantage by pre-adolescence on a selective visual attention task.
Several studies have demonstrated that early auditory deprivation (deafness) results in specific, compensatory changes in visual processing. In particular, deaf individuals exhibit enhanced performance for tasks performed in the visual periphery. Accordingly deaf individuals asked to make a key press in response to stimuli presented in the central or peripheral visual field, exhibit faster RTs than hearing individuals for peripheral targets but not for central ones
Whether this enhancement confers
The performance of deaf and hearing individuals on a computerized adaptation of the Useful Field of View task (UFOV;
The majority of studies reporting enhancement of visual attention to the periphery have recruited deaf individuals born deaf to deaf parents who learned American Sign Language (ASL) as a first language. This leaves open the possibility that enhancements in attention are restricted to this sub-population and do not generalize to the deaf population at large. This is of concern as most studies reporting deficient visual attention have focused on deaf non-signers. Therefore, in addition to recruiting deaf native signers, we included deaf individuals who experienced early auditory deprivation but did not learn a sign language. In addition, the impact of sign language was further evaluated using hearing subjects, born to deaf parents, who acquired ASL as a first language. Some of the studies referenced above have included hearing native signers, and have suggested that sign language use is not sufficient to induce enhanced peripheral attention
This research was approved by the Research Subjects Review Board at the University of Rochester, NY.
Each subject was assessed using a modification of the UFOV paradigm
A Schematic of Useful Field of View Task. In the experimental UFOV task, participants were asked to discriminate a briefly presented face in the center of the display – the cutaways show detail of the ‘short hair’ and ‘long hair’ faces – and to indicate the location of a peripheral target (a five-pointed star in a circle) via a touch screen. B Useful Field of View Thresholds, Experiment 1. Performance (mean threshold is ms) of each subject group on the experimental UFOV task; error bars indicate ±1 SEM.
All three tasks were presented within a circular gray field subtending 21° of visual angle. Each stimulus display was followed by a visual noise mask presented on the whole screen and then a prompt appearing at fixation. Participants indicated their response (in speech or sign) for the central task, and the experimenter manually entered that response. For the peripheral localization response, participants touched the screen at the location where they believed the peripheral target to have appeared. Trials were classified as correct if the subject accurately identified both the central icon's identity and the location of the peripheral target (in the first task, only central task performance applied). An adaptive staircase procedure was employed for all three tasks – after three consecutive correct responses, the stimulus duration was reduced by 1 frame (1/60 second); one incorrect response resulted in the stimulus duration being increased by 1 frame. Each task finished after twelve reversals, ten consecutive correct trials at ceiling (1 frame), or 72 trials, whichever occurred sooner. A threshold measure was calculated by averaging the stimulus duration of the last 10 correct trials. In the divided attention training and the UFOV tasks – which required both central and peripheral responses – trials where the central target was incorrectly identified were ignored (i.e. those trials were not used for computing step changes in the adaptive staircase procedure).
Stimuli were presented using Matlab software and the Psychophysics Toolbox installed on a Apple G4 Titanium laptop computer running OS 9.2.2. The laptop was connected to a 23″ Apple Cinema Display via an Apple ADC-DVI adaptor, with a 60 Hz refresh rate. The display was adapted to function as a touch screen using pressure-sensitive resistive (PSR-1®) technology, supplied and fitted by Troll Touch Touchscreens (Valencia, CA).
Subjects were tested in a single experimental session lasting approximately 25–30 minutes. Subjects were in a chin rest, positioned 40 centimeters from the center of the touch screen. Instructions were given in sign or speech and clarified if necessary. Subjects were given the correct answer on the first 2–3 trials if they still appeared to be confused.
All statistical tests were conducted with an α = .05. Confidence intervals for differences between group means (CI95diff) are reported alongside statistical test results and estimates of effect size (partial η2).
Potential adult subjects were asked about their videogame playing. Those who reported playing action-based videogames were classified as ‘game players’. This classification did not influence enrollment into the study, although data from ‘game players’ are not reported here as it is known that action video gaming changes performance on the UFOV
Deaf adult signers (N = 10, MAGE = 26.1, 2 males) were recruited at a school in Austin (TX) and at a camp in Madison (SD), as well as from participant pools at RIT/NTID (NY) and Gallaudet University (DC). All were deaf native signers who reported being born with severe-profound auditory deprivation (hearing loss >75 dB in the better ear; for 5 deaf signers who knew their exact level of hearing loss, mean loss in the better ear was 107 dB with a range of 75–120 dB) to deaf parents from whom they learned ASL as a first language. In the absence of a reliable and easily administered ASL proficiency test, subjects were asked to rate their ASL comprehension and production proficiency on a scale from 1 = perfect to 4 = hardly. All deaf signers gave themselves a rating of 1.0 in ASL comprehension and 1.0 in ASL production.
Deaf adult non-signers (N = 10, MAGE = 21.6, 3 males) were students recruited at the National Technical Institute for the Deaf (NTID) in Rochester, NY. All reported being born with severe-profound auditory deprivation (>75 DB in the better ear; for 6 deaf non-signers who knew their exact level of hearing loss, mean loss in the better ear was 90 dB with a range of 75–110 dB). Although most reported knowing some ASL, their first regular exposure to ASL had been at NTID where they were recruited for this study during their first quarter in order to limit that exposure. Accordingly, they reported an inability to communicate clearly in ASL (on average rating themselves 3.3 in ASL comprehension and 3.2 in ASL production). All deaf non-signing subjects preferred testing to be conducted using spoken English.
Hearing adult signers born to deaf parents (N = 10, MAGE = 22.9, 4 males) were recruited from a summer camp for KODAs (‘kids of deaf adults’) in Eagle Bay, NY. All reported learning ASL from their parents as infants, and expressed competence in ASL (on average rating themselves 1.8 in ASL comprehension and 1.8 in ASL production). None reported any hearing loss, and all testing was conducted in ASL.
Hearing adult non-signers (N = 10, MAGE = 20.4, 2 males) were recruited from a participant pool at the University of Rochester, NY. All reported normal hearing and no knowledge of any sign language.
Prior to analyzing the UFOV thresholds for the selective attention task (i.e. the task with distractors) it was important to establish that the central task was attentionally demanding in this context, and thus in competition with the peripheral target for attentional resources. While this task provided no independent, concurrent measure of central task performance, identification accuracy was calculated for the last 1/3 of trials for all subjects (see
UFOV thresholds (i.e. with distractors present) were entered into a two-way ANOVA with auditory deprivation (deaf, hearing) and signing (signer, non-signer) as between subjects factors (see
Although the two other tasks (central stimulus identification and divided attention) were included for training purposes, deaf non-signer participants differed from the other groups in a manner worthy of note (
Performance on the central training task (A) and central and abrupt peripheral onset training task (B) was generally asymptotic, except for deaf adults who did not use a signed language. For this group, the thresholds on these two tasks were significantly elevated. Error bars indicate ±1 SEM.
UFOV thresholds were reanalyzed with each subject's performance on these training tasks as covariates. The pattern of findings did not change, with the main effect of auditory deprivation remaining the sole significant effect (F(1,34) = 6.21, p = .018, partial η2 = .15, CI95diff = 4–39 ms).
This first experiment establishes the role of auditory deprivation in the enhancement of peripheral visual attention noted in the deaf population. Both deaf signing and deaf non-signing adults excelled at the UFOV task. This shows that the enhancement is not limited to the use of isolated targets but generalizes to complex tasks such as the UFOV, which combines selective visual attention with the requirements of performing two tasks (one centrally and the other peripherally). Although deaf non-signers displayed better performance on the UFOV task than their hearing peers, they showed worse performance on the central stimulus identification and divided attention tasks. This result is surprising in the face of their enhanced performance on the UFOV task. The two training tasks differ from the main UFOV task along several dimensions preventing us from drawing firm conclusions. The central identification task focuses entirely on central processing, rather than peripheral processing in the context of an additional central task like in the UFOV task. The divided attention task requires both peripheral and central processing in the same manner as for the UFOV task, but it differs from the UFOV task in terms of its very low attentional load
There are two alternative mechanisms that can be ruled out by the overall pattern of data reported. The first is that any deficits observed for deaf individuals stem from the need to make sequential manual responses (sign SHORT or LONG and then touch the screen) whereas hearing individuals can make a simultaneous oral-manual response (say “short” or “long” while touching the screen at the same time). If this were the case, then there should be a deficit for deaf signers across all tasks requiring two responses, which is clearly not the case. Despite the need to execute sequential responses for the two tasks, deaf signers outperform hearing subjects on the UFOV task, and show comparable performance on the divided attention task. Indeed, the deaf non-signers who performed poorly on the divided attention task made simultaneous oral and manual responses to the targets. The second alternative is a perceptual enhancement in the peripheral visual field of deaf individuals. Such an enhancement would predict enhanced performance on the divided attention task for all deaf individuals. To the contrary, deaf non-signers showed impaired performance on the divided attention task and deaf signers showed similar performance as their hearing peers. This pattern of finding reinforces the view that peripheral processing enhancements in deaf individuals result from changes in selective attention, and not perceptual modifications
In Experiment 2 we ask at what age such a redistribution of attention becomes apparent in a sample of deaf children compared to a group of hearing peers 7 to 17 years of age. Deaf children were recruited from a camp and deaf school where ASL was the primary means of communication. The experimental design, apparatus and procedure were the same as those employed in Experiment 1. Previous studies suggest that visual selective attention skills are relatively stable in hearing subjects by 7–10 years of age
Written informed consent was obtained from all children and a parent or legal guardian. All children were rewarded with a $15 gift card. As in Experiment 1, action video game players were tested but their data are not reported here.
Hearing children were recruited from a school district in the Rochester NY area. Mailings were sent from the school district to parents of all children aged 7 to 17 years. The response rate was approximately 15%. All children had normal or corrected-to-normal vision and no known history of neurological or cognitive impairment. They were screened to ensure none required an Individualized Educational Program (IEP) indicating the need for accommodations due to learning or language impairment. Children were divided into three age categories: 7–10 year old elementary/primary students (N = 38, MAGE = 9;1, 16 males), 11–13 year old middle school students (N = 16, MAGE = 12;2, 5 males), and 14–17 year old high school students (N = 14, MAGE = 15;7, 1 male).
Deaf children were recruited from deaf schools in Rochester NY and Austin TX, and a camp for deaf children in Madison, SD. School or camp administrators mailed letters to the parents of all children aged between 7 and 17 years, resulting in a 10% response rate. All children had normal or corrected-to-normal vision and no known history of neurological or cognitive impairment. Although most of the deaf children had IEPs as a result of their deafness, none had reported attentional problems or learning disabilities. The deaf children divided into the same age categories as hearing children: 7–10 year olds (N = 15, MAGE = 9;3, 10 males), 11–13 year olds (N = 20, MAGE = 12;4, 1 male), and 14–17 year olds (N = 14, MAGE = 15;6, 7 males). All had an unaided hearing loss >70 dB in their better ear and used ASL on a daily basis as their primary means of communication. None had undergone cochlear implant surgery. Sixteen (33%) had hearing parents, although all of these children had started to learn ASL in pre-K classes. Parental hearing status had no effect on the measures used, and is not considered further. Given their background, this group is more similar to the native signers adults described above, and differs in aetiology from the children typically considered in the literature on deafness, visual attention and cochlear implants
A two-way ANOVA on experimental UFOV thresholds with auditory deprivation (deaf, hearing) and age group (7–10, 11–13, 14–17 years) as between subjects factors revealed significant main effects of auditory deprivation (F(1,117) = 17.85, p<.001, partial η2 = .14, CI95diff = 11–31 ms) and age group (F(2,117) = 6.49, p = .002, partial η2 = .11), and a significant two-way interaction between auditory deprivation and age group (F(2,117) = 7.10, p = .001, partial η2 = .11). This interaction led us to assess the effects of age group separately for deaf and hearing children. As predicted, for hearing children the UFOV thresholds did not vary as a function of age group (F(2,68) = 0.12, p = .884, partial η2<.01), whereas they did for deaf children (F(2,49) = 25.43, p<.001, partial η2 = .53). While deaf 7–10 year olds performed equivalently to hearing 7–10 year olds, older deaf children demonstrated better thresholds, outperforming their hearing peers and the youngest deaf children (see
In the main UFOV task the performance of 7–10 year old deaf children was comparable with that of their hearing peers, whereas older deaf children were significantly better than their hearing peers. Error bars indicate ±1 SEM.
Interestingly, the training tasks indicated worse performance in the youngest deaf group compared to the other groups (see
On the two training tasks, the youngest deaf children (7–10 year olds) performing significantly worse than their hearing peers. Across all age ranges tested, for both deaf and hearing samples, these children were the only ones who did not perform near ceiling on these tasks.
The UFOV task requires subjects to divide attention between central and peripheral locations, while also selecting a target from amongst distractors. It is an attention-demanding task, requiring not only central attention but also attention to the periphery
Although a tendency for more effective visual search in deaf than in hearing individuals has been reported
Data from children revealed that this attentional enhancement is not observed until after 7–10 years of age, although the precise point within this age group could not be determined due to sample size limitations. Nevertheless, there is the suggestion that a robust cross-modal enhancement in visual selective attention is not observed until after several years of auditory loss. Further study is required to identify exactly when and how this delayed enhancement is brought about. The lack of improvement observed in hearing children suggests that maturational factors are unlikely to contribute. Rather, it may be the duration of auditory deprivation that plays the key role, with over 7–10 years of auditory deprivation required for effects to be manifested behaviorally. Alternatively, it may reflect a ‘sleeper effect’
There is some evidence that the youngest deaf children found the central stimulus identification task to be more difficult than did their hearing peers, with this difficulty extending to the divided attention task. Data from deaf adults suggest that any such deficit is no longer apparent by adulthood, at least for those who have early and full access to a first language (deaf native signers). All of the young deaf children in this study had early language access through ASL, although the extent of their social and linguistic interactions with caregivers during early infancy cannot be assessed post hoc. It is important to note that studies reporting deficiencies in visual attention skills have typically used central visual field tasks employing rapid stimulus presentations with young deaf children
This work establishes that auditory deprivation is not a causal factor for attentional difficulties. All deaf individuals tested performed at least as well and often significantly better than their hearing peers on the UFOV measure, an attentionally-demanding task Worse performance in the youngest deaf children and those deaf adults with limited access to a natural language early in development was noted under some conditions. While these results are in line with previous work documenting attentional deficits in deaf children, the present study makes it clear that such challenges early in childhood are not predictive of deficient functioning as development proceeds.
Central task performance in selective UFOV task. For the UFOV selective attention task, mean central identification accuracies and mean stimulus presentation durations were calculated based upon the last 1/3 of trials for each subject. Accuracy levels indicate that the central task is attentionally demanding for all subject groups. However, accuracy cannot be compared directly across groups, as presentation durations differed. After normalizing accuracies as a function of presentation duration, performance did not significantly differ as a result of deafness or sign language use.
(0.04 MB DOC)
Two-alternative forced choice (2-AFC) discrimination task at the center of the visual field - a face icon was presented in the center of the screen and participants had to decide whether it had long or short hair.
(0.42 MB MOV)
The same 2-AFC central discrimination task as in
(0.62 MB MOV)
The 2-AFC central discrimination task with localization of a peripheral target presented at 20° of visual angle at one of eight possible cardinal/intercardinal locations and embedded in a field of distractors. The distractors appeared along the eight possible cardinal/intercardinal axes at 6.67, 13.33 and 20 degrees of visual angle.
(0.52 MB MOV)
Thanks to Dara Baril and Wyatte Hall for help with data collection, and to the following for their enthusiastic participation in the study: the staff and students of Texas School for the Deaf (Austin, TX), Camp Lakodia (Madison, SD), Rochester School for the Deaf (Rochester, NY), Brighton Central School District (Rochester, NY); and Camp Mark Seven (Eagle Bay, NY). We would also like to thank our reviewers for their insightful comments and help in improving the manuscript.