1 23

Attention, Perception, &Psychophysics ISSN 1943-3921 Atten Percept PsychophysDOI 10.3758/s13414-014-0695-2

The role of head movements in thediscrimination of 2-D shape by blindecholocation experts

Jennifer L. Milne, Melvyn A. Goodale &Lore Thaler

1 23

Your article is protected by copyright

and all rights are held exclusively by The

Psychonomic Society, Inc.. This e-offprint is

for personal use only and shall not be self-

archived in electronic repositories. If you wish

to self-archive your article, please use the

accepted manuscript version for posting on

your own website. You may further deposit

the accepted manuscript version in any

repository, provided it is only made publicly

available 12 months after official publication

or later and provided acknowledgement is

given to the original source of publication

and a link is inserted to the published article

on Springer's website. The link must be

accompanied by the following text: "The final

publication is available at link.springer.com”.

The role of head movements in the discrimination of 2-D shapeby blind echolocation experts

Jennifer L. Milne & Melvyn A. Goodale & Lore Thaler

# The Psychonomic Society, Inc. 2014

Abstract Similar to certain bats and dolphins, some blindhumans can use sound echoes to perceive their silent sur-roundings. By producing an auditory signal (e.g., a tongueclick) and listening to the returning echoes, these individualscan obtain information about their environment, such as thesize, distance, and density of objects. Past research has alsohinted at the possibility that blind individuals may be able touse echolocation to gather information about 2-D surfaceshape, with definite results pending. Thus, here we investigat-ed people’s ability to use echolocation to identify the 2-Dshape (contour) of objects. We also investigated the roleplayed by head movements—that is, exploratory movementsof the head while echolocating—because anecdotal evidencesuggests that head movements might be beneficial for shapeidentification. To this end, we compared the performance ofsix expert echolocators to that of ten blind nonecholocatorsand ten blindfolded sighted controls in a shape identificationtask, with and without head movements. We found that theexpert echolocators could use echoes to determine the shapesof the objects with exceptional accuracy when they wereallowed to make head movements, but that their performancedropped to chance level when they had to remain still. Neitherblind nor blindfolded sighted controls performed abovechance, regardless of head movements. Our results show notonly that experts can use echolocation to successfully identify

2-D shape, but also that head movements made whileecholocating are necessary for the correct identificationof 2-D shape.

Keywords 2-D shape and form . Audition . Hearing

It is well known that some animals use self-generated soundsto perceive their surroundings via reflected sound waves, orechoes. Echolocation can be used in environments not condu-cive to vision, thereby allowing animals to navigate andforage even in complete darkness. Similarly, some blindhumans have developed the ability to use echoes from self-produced sounds to perceive their silent surroundings. Forexample, blind echolocators can perceive information suchas the size, shape, distance, motion, and material properties ofsilent objects (Arnott, Thaler, Milne, Kish, & Goodale, 2013;Kellogg, 1962; Rice, 1967, 1969; Rice & Feinstein, 1965;Rice, Feinstein, & Schusterman, 1965; Schenkman &Nilsson, 2010; Stoffregen & Pittenger, 1995; Teng, Puri, &Whitney, 2012; Teng & Whitney, 2011; Thaler, Arnott, &Goodale, 2011; Thaler, Milne, Arnott, Kish, & Goodale,2013). In this way, then, echolocation could be considered acrude substitute for vision, allowing blind humans to perceiveaspects of their environment that would otherwise goundetected.

Although our knowledge of echolocating animals such asdolphins and some species of bats is quite extensive (see, e.g.,Harley, Putman, & Roitblat, 2003; Schnitzler & Kalko, 2001;Thomas, Moss, & Vater, 2004), comparably little research hasbeen dedicated to understanding the use of echolocation byhumans. In the 1940s, it was determined that blind individ-uals’ ability to avoid obstacles and sense the presence ofobjects was not due to “facial vision,” but to the use of activeauditory perception (Ammons, Worchel, & Dallenbach, 1953;Cotzin & Dallenbach, 1950; Supa, Cotzin, & Dallenbach,

J. L. Milne :M. A. GoodaleBrain and Mind Institute, University of Western Ontario, London,Ontario, Canada

L. ThalerDepartment of Psychology, Durham University, Durham, UK

M. A. Goodale (*)Brain and Mind Institute, University of Western Ontario, London,Ontario, Canada N6A 5B7e-mail: mgoodale@uwo.ca

Atten Percept PsychophysDOI 10.3758/s13414-014-0695-2

Author's personal copy

1944). Following this discovery, a series of behavioral inves-tigations of echolocation revealed the ability of blindecholocators to detect the presence of objects, and also tocomment on object features such as size, distance, and mate-rial properties (Kellogg, 1962; Rice, 1967; Rice et al. 1965;Schenkman & Nilsson, 2010; Schörnich et al., 2013; Tenget al., 2012). Studies have even provided evidence that sightedindividuals can learn to echolocate, as well (Ammons et al.,1953; Teng & Whitney, 2011).

Although human echolocation is receiving increasing at-tention in the literature, a clear understanding of the ability ofblind echolocators to discern 2-D shape is lacking. The per-ception of shape is likely important to a blind echolocator—for example, during navigation, for which landmark identifi-cation and obstacle avoidance are critical. In 1967, Ricereported preliminary results of a 2-D shape discriminationtask that suggested that blind echolocators could distinguishbetween a circle, square, and triangle, but he never followedup on these initial observations. Later, Hausfeld, Power,Gorta, and Harris (1982) showed that untrained sighted indi-viduals could learn to discriminate simple shapes using ech-oes, and that a blind participant performed within the range ofthese sighted individuals. Furthermore, we know from theliterature that echolocating bats can perceive the shapes ofobjects from echoes and can use this information to discrim-inate between food and nonfood objects (Simmons & Chen,1989). Thus, it is reasonable to believe that blind expertecholocators can determine the shape of objects using echoes.What remains unclear in the literature, though, is how the useof movement affects shape identification. We know, anecdot-ally, that when expert echolocators are naturally using echo-location, they typically make many movements with theirhead. In fact, almost all of the studies on echolocation fromthe last century mentioned that their echolocating participantsused head movements that seemed to aid in performance onthe tasks, but no research has been done to follow up on thesereports. In the context of 2-D shape perception, head move-ments are likely to be useful for resolving the 2-D shape of anobject—for example, by acoustically “tracing the contour” ofan object.

In sum, on the basis of the evidence to date, the aim of thepresent study was twofold: (1) to determine whether blindexpert echolocators can use echolocation to identify 2-Dshapes, and (2) to determine whether this behavior is affectedby imposing constraints on their head movements. We foundthat expert echolocators were remarkably accurate at identify-ing shapes when they were allowed to freely move and ex-plore the objects as they would naturally; when they wererequired to remain still, however, their performance declineddramatically. The results of the present work contribute to ourunderstanding of the applications of echolocation and showthat head movement is crucial to successful object identifica-tion (at least in the case of 2-D shape perception).

Method

Participants

A total of 26 participants were recruited to participate in ashape identification experiment at the University of WesternOntario (London, Ontario, Canada). All testing procedureswere approved by the university ethics board, and participantsgave written informed consent prior to testing. Participantswere drawn from three different groups: blind expertecholocators (EE; who reported everyday use of active echo-location and extensive experience with the technique), blindcontrols (BC; who reported little to no use of active echolo-cation techniques), and blindfolded sighted controls (SC; whoreported normal or corrected-to-normal vision and no experi-ence with echolocation techniques). Blind participants whoreported any residual vision (e.g., bright light detection) werealso blindfolded. All participants reported having normalhearing and no history of hearing difficulties. See Table 1for the participant details.

It is important to note that blind and blindfolded sightedcontrols had received no echolocation training prior to partic-ipating in the experiment. It is clear from previous researchthat sighted individuals can learn to use echoes (Teng &Whitney, 2011), and of course, blind individuals can betrained as well. The purpose of the control participants in thepresent study, however, was to control for performance thatcould be attributed to factors other than echolocation expertise(super-sensitivity to echoes as a simple consequence of blind-ness, ambient sounds, sounds from the movements of theexperimenter, etc.). The tongue click, finger snap, and otherecholocation signals were explained to control participants,and they were free to use the technique of their choosing,provided that the “signal” was produced without any externaldevice.

Stimuli

Four two-dimensional shapes were presented to all partici-pants: a square (40 × 40 cm), a triangle1 (52 × 45 cm [height]),a rectangle oriented horizontally (100 × 16 cm), and the samerectangle oriented vertically (see Fig. 1a). All of the shapeswere made of a 0.5-cm-thick foam board and covered withaluminum foil. The shapes were positioned on a 0.6-cm-diameter pole, which was determined to be undetectable by

1 Please note that the surface area of the triangle is slightly smaller thanthose of the other stimuli. This was done to make the triangle “visuallysimilar” in size to the other shapes. We had participants EE1–EE3 verifythat this difference in surface area would not be informative, regardless ofshape. Furthermore, as we will also mention in the Results, we confirmedthrough an analysis of the error distributions that performance was thesame in the “triangle” conditions as for the other shapes.

Atten Percept Psychophys

Author's personal copy

echolocation. Before beginning the experiment, all partici-pants were familiarized with the four shapes (sighted controlswere allowed only to touch the shapes and not to see them).

Procedure

All participants took part in two conditions: a “free-moving”condition, permitting head and body movements, and a“fixed-position” condition, not permitting any movement.For both conditions, participants EE1, EE2, and EE5 weretested in the Beltone Anechoic Chamber (18 feet high, 23 feetwide, 12 feet deep) at the National Centre for Audiology inLondon, Ontario, Canada. The chamber is equipped with a125-Hz cutoff wedge system on the walls and ceiling, andambient noise recordings indicated a noise floor of 18.6 dB(Larson-Davis System 824). Only participants EE1, EE2, andEE5 were tested in the anechoic chamber due to logisticalreasons (i.e., additional participants were not available at thetime of testing, and the researchers had limited access to thechamber). For the free-moving conditions, these participantswere also tested in an echo-dampened room (2.75 × 3 m, four

walls covered in 3.8-cm convoluted foam sheets). After de-termining that there were no performance differences betweenthe anechoic chamber and the echo-dampened room (see theResults), we felt that it was not necessary to test other partic-ipants in the anechoic environment. Therefore, all other par-ticipants in all conditions were tested in the echo-dampenedroom only.

On each trial, one of the four shapes was presented. Thepresentation height was unique for each participant, in order tocenter the shapes at ear level. For the free-moving conditions(Fig. 1b), participants were situated at a starting position40 cm away (measured from the ears) and centered on theshape. Once the trial began, participants could freely movetheir heads and/or bodies to examine the objects via echolo-cation. For the fixed-position condition (Fig. 1c), participantswere situated 80 cm away from the shape and had to keep theirhead and body still for the duration of the trial. This fartherdistance (relative to the 40-cm starting distance in the free-moving condition) was reported by three expert echolocatorsas providing the “best overall impression” of the shape. Theymentioned that being any closer to the objects in the fixed-

Table 1 Participant details for expert echolocators (EE), blind controls (BC), and blindfolded sighted controls (SC)

Group Participant Sex, Age Cause of Blindness Onset Residual Vision Echolocation Technique

EE EE1 M, 45 Retinoblastoma Early None Tongue click

EE2 M, 29 Glaucoma Early, progressive None Tongue click

EE3 M, 56 Optic nerve atrophy Early None Tongue click

EE4 M, 49 Retinoblastoma Early None Speech, tongue click

EE5 M, 44 Retinopathy of prematurity Early None Tongue click

EE6 F, 21 Idiopathic intracranial hypertension Late Bright light Finger snap

BC BC1 M, 39 Leber congenital amaurosis Early, progressive Bright light Speech, finger snap

BC2 M, 34 Retinopathy of prematurity Early Bright light (left eye) Finger snap, clap

BC3 F, 25 Diabetes Late Low level vision (left eye) Finger snap, clap

BC4 F, 22 Glaucoma, cataracts Early Bright light Speech, clap, finger snap

BC5 M, 44 Retinopathy of prematurity Early, progressive Bright light Speech, finger snap

BC6 M, 20 Leber congenital amaurosis Early, progressive Bright light (left eye) Finger snap

BC7 M,40 Optic nerve atrophy Early None Clap, finger snap

BC8 F, 60 Retinoblastoma Early None Speech, clap

BC9 F, 24 Retinopathy of prematurity Early Bright light Clap

BC10 F, 32 Optic nerve atrophy Early Low-level vision Clap

SC SC1 M, 30 Clap, speech

SC2 M, 29 Clap, finger snap

SC3 F, 22 Clap

SC4 F, 58 Clap

SC5 F, 31 Clap

SC6 F, 45 Clap, finger snap

SC7 F, 47 Clap

SC8 F, 37 Clap

SC9 F, 35 Clap

SC10 F, 56 Clap, finger snap

Atten Percept Psychophys

Author's personal copy

position condition would prevent them from gathering objectedge information from the echoes. The 80-cm position wasused for the fixed-position condition, then, to provide the bestpossible chance for successful performance in these cases. Tovalidate the suggestion given by the expert echolocators, andto rule out distance as a confound, we conducted a controlexperiment that replicated the “fixed-position” conditions, butat a distance of 40 cm (Fig. 1d). This experiment was con-ducted only with a subset of the participants (EE3, EE6, SC2,and SC3).

Throughout the experiment, participants used an echoloca-tion technique of their choice (see Table 1) and listened forreflected echoes to determine the shape of the stimulus pre-sented. For the fixed-position condition, any participants whochose an echolocation technique other than tongue clicks orother vocalizations were asked to keep the source of the sound(e.g., their hand while finger snapping) underneath the chinand as close to the body as possible, and they could not movefrom that position. For both conditions, participants weregiven a maximum of 15 s per trial and could provide theirresponse at any point within that time frame (four-alternativeforced choice—“square,” “triangle,” “horizontal rectangle,”or “vertical rectangle”). For each trial, the experimenter mea-sured the participants’ response times (i.e., trial start untilverbal response onset) using a stopwatch. For each con-dition, a total of 40 pseudorandom trials were presented(ten repetitions per shape, per condition).

Results

For the purpose of the analyses, performance for each partic-ipant was collapsed across the four shapes (analyses notshown here revealed no significant differences in the individ-ual shape response patterns for any of the groups). Therefore,the analyses were performed on the overall percentage correctvalue for each participant in each of the conditions (free-moving and fixed-position). As was mentioned in theMethod section, for the free-moving condition participantsEE1, EE2, and EE5 were tested in both an anechoic chamberand an echo-dampened room. For each of these participants,we ran t tests to determine whether any performance differ-ences would emerge between the anechoic chamber and theecho-dampened room. The results for all three participantsrevealed no significant differences in performance betweenthe two rooms [EE1, t(78) = –5.30, p = .598; EE2, t(78) =1.63, p = .107; EE5, t(78) = –1.113, p = .269]. Therefore, forthe purpose of the following analyses, we averaged theseparticipants’ performance scores across the two testingenvironments.

A 3 × 2 mixed analysis of variance (ANOVA) was con-ducted on the data, with the between-subjects factor Groupand the within-subjects factor Condition. The Group factorincluded three levels: expert echolocators (n = 6), blind con-trols (n = 10), and sighted blindfolded controls (n = 10). TheCondition factor included two levels: free moving and fixedposition. Because fewer participants were in the EE group,and therefore variability could differ across the groups, wecomputed Levene’s tests for each condition. The results forboth conditions were not significant [free-moving, F(2, 23) =2.371, p = .116; fixed position, F(2, 23) = 2.61, p = .095].

The results of the ANOVA revealed a significantinteraction between condition and group, F(2, 23) = 38.535,p < .0005, η2 = .77 (see Fig. 2a). Bonferroni-corrected pairwisecomparisons revealed that the EE group performed significant-ly better than both the blind (p < .0005) and blindfoldedsighted (p < .0005) control groups in the free-moving condi-tion. In the fixed-position condition, the EE group also per-formed significantly better than both of the control groups (EEvs. BC, p = .012; EE vs. SC, p = .043), but this difference wassubstantially smaller (EE vs. BC, mean difference = 14.83; EEvs. SC, mean difference = 12.33). The drastic decline in the EEgroups’ performance in the fixed-position condition is easilyseen in Fig. 2a, and pairwise comparisons revealed that the EEgroup performed significantly better in the free-moving con-dition (p < .0005). The performance of the two control groupsin both conditions was statistically indistinguishable (freemoving, p = 1; fixed position, p = 1). Overall, the results ofthe interaction show that when the expert echolocators couldfreely moving their heads and bodies, they had a substantialadvantage and were able to reliably indicate the shape of theobject presented to them. When they were required to remain

Fig. 1 Stimuli and procedure for the free-moving and fixed-positionconditions. Four 2-D shapes (a) were presented individually to partici-pants. The shapes were made of foam board and covered in aluminum foilto maximize sound reflection. In the free-moving condition (b), partici-pants were situated 40 cm away from the shape, which was centered at earlevel. Once the trial began, participants could move freely in any direction(without touching the shape) in order to identify the shape via echoloca-tion. In the fixed-position condition (c), participants were situated 80 cmaway from the shape and had to remain in that position for the duration ofthe trial, without making any movements. We also ran a control experi-ment (d) on a subset of the participants, to rule out the possibility of adistance confound between the free-moving and fixed-position condi-tions. For all conditions, participants were given a maximum of 15 s pertrial to identify the presented shape

Atten Percept Psychophys

Author's personal copy

still, however, their ability to indicate the shape of the objectsdecreased dramatically. Neither of the control groups showedthis movement advantage.

To address the possibility that the distance differences perse were responsible for the decrease in performance in thefixed-position as compared to the free-moving condition, weconducted a control experiment with a subset of the partici-pants. This control experiment replicated the fixed-positioncondition, but at the 40-cm position. We used nonparametricrelated-samples McNemar’s tests to compare the individualparticipants’ performance on fixed-position conditions at eachof the two distances. The results of this experiment are shownin Fig. 2b. The data show no advantage to being located at40 cm in the fixed-position conditions. In fact, the EE partic-ipants’ performance was the same as (EE3, p = 1) or worsethan (EE6, p = .008) their performance in the 80-cm fixedposition conditions. The sighted controls showed no signifi-cant difference in performance between the two distances(SC2, p = .5; SC3, p = .25). In sum, the difference in objectdistance between free-moving and fixed-position conditions

could not account for the performance differences in the EEgroup.

Main effects of both group and condition were also found[F(2, 23) = 42.189, p < .0005, η2 = .786; F(1, 23) = 46.637,p < .0005, η2 = .67]. Bonferroni-corrected pairwise compari-sons for the main effect of group revealed that the EE groupperformed significantly better than both of the control groups(EE vs. BC, p < .0005; EE vs. SC, p < .0005), but that thecontrol groups performed identically (p = 1). Inspection of themeans showed that the main effect of condition was due to thefact that, overall, participants performed significantly better inthe free-moving condition than in the fixed-position condition(p < .0005). This effect, of course, was driven by the highperformance of the EE group in the free-moving condition, aswas shown in the significant interaction.

To supplement the ANOVA analysis, we also ran individ-ual t tests on each group for each condition, comparing per-formance to chance (25 %), and the results were Bonferronicorrected for multiple comparisons. Performance was signif-icantly different from chance only for the EE group in the free-

Fig. 2 Percent correct performance of all groups (expert echolocators[EE], blind controls [BC], and sighted blindfolded controls [SC]) in eachof the two conditions (free-moving and fixed-position). Panel A presentsthe results of the omnibus ANOVA, which revealed a significant interac-tion between the factors (significant differences are indicated by aster-isks). Error bars represent the standard errors of the means across partic-ipants, and the dashed line indicates chance performance (25 %). Panel Bshows the performance of two expert echolocators and two sightedcontrols, who were also tested at the 40-cm position in the fixed-positioncondition. For reference, we also show those participants’ performance at

the 80-cm fixed-position condition. It is clear from the data that beinglocated closer to the objects in the fixed-position condition provided noadvantage. In fact, the echolocating participants showed comparable(EE3) or worse (EE6) performance at this distance (asterisks indicatesignificant differences in performance at the two distances, on the basis ofresults from nonparametric related-samplesMcNemar’s tests). These datasupport our use of the 80-cm position in this condition, which, accordingto the expert echolocators, provided the best overall impression of theshape, and thus a better chance of successful performance

Atten Percept Psychophys

Author's personal copy

moving condition, t(5) = 8.013, p < .0005. The EE group didnot perform significantly better than chance when they wererequired to remain still [t(5) = 2.019, p = .099], and the BC andSC groups performed at chance level in both the free-movingand fixed-position conditions [BC-free, t(9) = 0.023, p = .982;BC-fixed, t(9) = –0.943, p = .370; SC-free, t(9) = –0.103,p = .920; SC-fixed, t(9) = 0.514, p = .619]. The results of thetests against chance are consistent with the ANOVA, in thatthey provide support for a strong advantage for the EE groupsin the free-moving condition.

Although the ANOVA allowed us to gain an understandingof the overall performance of the EE group as compared to thecontrol participants, it is important to appreciate that, similarto neuropsychological patients, blind echolocators show pro-found variability in their echolocation abilities as well as intheir histories of use, causes and times of blindness, and so on.Therefore, we felt that it was important to also analyze the databy treating each individual echolocator as a single case andcomparing their performance in both of the conditions to thatof the control participants. To increase statistical power forthis analysis, and because the ANOVA revealed no significantdifferences in performance between the control groups, wecombined the control groups for each condition (free movingand fixed position) for the purposes of this analysis. For eachEE participant, we ran modified t tests to compare theirperformance to that of the combined control group for bothconditions (see Crawford & Garthwaite, 2002, 2007, 2012;Crawford, Garthwaite, & Porter, 2010; Crawford & Howell,1998). The modified t test is an extension of the traditional ttest, but it has been adapted to compare a single case to acontrol group. For the free-moving condition, the results of themodified t tests revealed that each individual echolocatingparticipant performed significantly better than the combinedcontrol group (see Table 2 for all t and p values, and Fig. 3a fora graphical depiction of each individual’s performance againstthe control group). The effect size of each echolocator’s scorealso reveals that they reliably performed well above the level

of the control group (see Fig. 3b). In the fixed-position con-dition, however, only participants EE2, EE4, and EE6 per-formed significantly better than the control group, and theeffect size of the difference in performance was substantiallylower than in the free-moving condition (with the exception ofEE6, who showed high performance in both conditions; seeFig. 3).

We also ran a 3 (group) × 2 (condition) mixed ANOVA onthe participants’ response times, but this analysis did notreveal any significant results [Condition × Group, F(2, 23) =0.524, p = .599, η2 = .044; condition,F(1, 23) = 4.14, p = .054,η2 = .153; group, F(2, 23) = 2.995, p = .07, η2 = .207]. Weobserved a trend toward significance for the main effect ofcondition, suggesting that, overall, participants used slightlymore time in the free-moving condition, but Bonferroni-corrected pairwise comparisons did not reveal a significantdifference.

Overall, the results show that expert echolocators can con-sistently and reliably indicate the shapes of 2-D objects whenthey are allowed to make head and body movements whileecholocating. When they are required to remain still, however,performance drops to a level that is statistically indistinguish-able from chance. In fact, our single-case analysis showed thatin free-moving conditions, all of our experts had performancestatistically superior to that in the control group, whereas onlyhalf of them maintained this superior performance in fixedconditions. Neither blind nor sighted blindfolded controlsshowed a movement advantage in the free-moving condition;in fact, their performance was nearly identical in each of theconditions, and never deviated from chance.

Discussion

The aim of the present experiments was to determine (1)whether blind expert echolocators can determine the 2-Dshape of objects by analyzing the echoes reflected from the

Table 2 Results of a modified t test analysis comparing individual echolocators to nonecholocating controls

Free-Moving Fixed Position

Control Sample Significance Test Control Sample Significance Test

n Mean SD Case Score t p n Mean SD Case Score t p

20 24.94 6.3 EE1 77.5 8.14 .000 20 24.75 5.61 EE1 27.5 0.478 .319

EE2 86.25 9.495 .000 EE2 37.5 2.217 .019

EE3 85 9.302 .000 EE3 27.5 0.478 .319

EE4 77.5 8.14 .000 EE4 37.5 2.217 .019

EE5 48.75 3.687 .000 EE5 30 0.913 .186

EE6 97.5 11.238 .000 EE6 70 7.866 .000

Means (control groups only) and case scores are percentage values (percent correct performance). Significance values (p) are one-tailed.

Atten Percept Psychophys

Author's personal copy

edges of similar objects, and (2) whether movements of thehead and body while echolocating are crucial for successfulshape identification in our task. The results were clear: Expertecholocators were exceptional at determining the shapes ofobjects that differed only in their edge or contour properties,and they performed well above the levels of blind participantswho do not echolocate and sighted participants who wereblindfolded. When the echolocators were required to remainstill, however, performance fell substantially, and wasstatistically indistinguishable from chancel level. Therefore,our results show that blind expert echolocators can use echoesto successfully determine the shapes of similar objects, andthat this ability is critically dependent on the use of headmovements. In our study, echolocators could move freely,which means that they could perform both angularmovements and movements in depth. Future research shouldaim to investigate the relative contributions of these separateaspects of head motion in more detail.

As we mentioned in the Method section, the blind andsighted control participants did not receive any explicit echo-location training prior to participation. Not surprisingly, then,these participants were unable to successfully use echoes todiscern the shapes of the objects. This runs contrary to thefindings of Hausfeld et al. (1982), who found that untrainedsighted participants could identify simple shapes using

echoes. Although these participants were untrained, they re-ceived feedback on every trial, and improvements in perfor-mance over the first few trials indicated that this feedback wasuseful. In fact, the participants in Hausfeld et al.’s studyreported that during the initial trials they were simply memo-rizing which echo was associated with which shape, and thenthey applied this knowledge to the remaining trials. It isunclear, then, whether the participants were actually perceiv-ing object shape or were simply relying on subtle differencesin echo characteristics, without perceiving any shape details.Therefore, the role of feedback and other methodologicaldifferences may explain the differences in performance be-tween untrained participants in the present study and inHausfeld et al.’s experiments.

An important consideration in the design of the experimentis the fact that the distances at which the shapes were present-ed were different for the free-moving and fixed-position con-ditions. Therefore, one could argue that the difference indistance alone might underlie the EE group’s decrease inperformance in the fixed-position condition. We addressedthis question in our control experiment, and the results of thatexperiment (see Fig. 2b) suggest that distance per se cannotexplain the performance differences between the free-movingand fixed conditions. Furthermore, if it were the case thatdistance per se could account for performance differences

Fig. 3 Results of the individual-case analyses for the free-moving andfixed-position conditions. Each individual echolocator’s performance inthe free-moving (y-axis) and fixed-position (x-axis) conditions is shownin Panel A. The data from the combined control group are also shown,with the shaded bars in each direction indicating the range of scores foreach condition. Significant results from the modified t tests are indicatedby asterisks. Asterisks above a data point indicate a significant differencefrom the combined control group in the free-moving condition, andasterisks to the right of a data point indicate a significant difference fromthe combined control group in the fixed-position condition (see Table 2

for the results from all individual tests). Dashed lines in each directionrepresent chance performance. The Bayesian effect size (with error barsshowing 95 % confidence intervals [CIs]; in some cases, the CIs are sosmall that error bars are not visible) for the results of each individual t testare shown in Panel B. The effect size was calculated using adapted zscores (Crawford et al., 2010). The “abnormality” of the case’s scores arepresented in Panel C, which shows the percentage of the control popula-tion (with 95%CIs) that would have obtained a lower score than the case.The information presented here and in Table 2 fully meets the reportingstandards set out by Crawford et al. (2010)

Atten Percept Psychophys

Author's personal copy

between the free-moving and fixed-position conditions, wewould expect to find a similar distance effect for all groups,but this was not the case. Finally, we want to highlight oncemore that the farther position for the fixed-position conditionwas chosen on the basis of echolocators’ advice, because theyfound that this distance gave them a better impression ofshape, as compared to closer distances. This can be under-stood, considering that if an individual is situated very close toan object and is required to remain still, the majority of theecholocation signal will be reflected from the center of theobject, and thus will lack edge information that could be usedto discern the object’s shape. One could imagine a similarsituation in vision, when an individual is situated very close toan object and is unable to gather information about objectfeatures in the periphery without movements of the head and/or eyes. This problem could be solved by simply movingfarther away from the object.

A final thing to consider is that more people in the blindcontrol group reported having some residual vision than did theexpert echolocators (see Table 1). It is possible that the presenceof some residual vision in a blind individual might make themless inclined to develop echolocation as a strategy. But thisneed not always be the case. Participant EE6, for example, hadsome residual vision at the time of testing, but even so hadmastered echolocation and performed better on the task thanany of the other expert echolocators. In any case, it seemsunlikely that the degree of vision normally available determineshow well people can use echoes to discriminate shape. Afterall, the blind controls did not perform better than the sightedcontrols when both groups were blindfolded. Furthermore, thetwo totally blind individuals (BC7, BC8) in the blind controlgroup performed no better or worse than the rest of that controlgroup, again suggesting that expertise per se, and not the degreeof blindness, drove performance in our study.

In sum, our results show that blind expert echolocators canuse echoes to successfully determine the shapes of similarobjects, and this ability is critically dependent on the use ofhead movements.

Head movements made while echolocating may be similarto the multiple eye movements, or saccades, that a sightedperson makes when visually scanning a large object or ascene. These saccades allow a person to accumulate visualinformation from the boundaries of a large object and thefeatures of a visual scene, which are then pieced together tocreate an overall perceptual representation. This process,termed transsaccadic integration, requires the brain to makequick computations of the incoming visual information inorder to arrive at a rich and stable representation of an objector scene (Niemeier, Crawford, & Tweed, 2003; Prime, Vesia,& Crawford, 2011). In terms of echolocation in the presentstudy, making headmovements while producing tongue clicks(or other signals) could have provided sound snapshots—or“echo saccades”—that are then automatically pieced together

by the brain to provide the individual with a perceptual repre-sentation of the object. Although transsaccadic integration invision can occur in a few hundred milliseconds, human echo-location is by comparison much more time-consuming andeffortful. Furthermore, the resulting percepts are likely coarserthan in vision. In fact, it has recently been shown that theprecision of echolocation is comparable to visual acuity in theperiphery, which, when compared to foveal acuity, is quitepoor (Teng et al., 2012).

Further evidence to support our suggestion that the headmovements made by echolocating humans might serve afunction similar to that of visual saccades comes from a recentstudy on scanning movements in echolocating bats (Seibert,Koblitz, Denzinger, & Schnitzler, 2013). It has been suggestedthat each bat signal–echo pair is comparable to a visualfixation and that the movements made by the bat betweensignal–echo pairs are comparable to visual saccades. Theresearchers found that the bats’ scanning behaviors changeddepending on the environment they were in and the task thatthey were performing. In particular, when the bats were ex-amining a scene, they made large scanning movements, butwhen they detected an object or obstacle, the angle of themovements was much smaller. This is quite similar to vision,in that the pattern of head and eye movements can be quitedifferent, depending onwhether one is looking at a large visualscene—which requires larger, longer movements—or at anobject within a visual scene—which requires smaller, shortermovements to gain greater object-specific detail (Hardiess,Gillner, & Mallot, 2008; Rayner, 1998). Considering boththese findings in echolocating bats and our results showingthe advantage of using head movements in human echoloca-tion, it will be important for future research to address thedifferent types of movements made by expert echolocators andhow these movements change in different environments andtasks.

Although it is quite clear that head movements—or “echosaccades”—seem to facilitate 2-D shape perception in echo-location, one of our echolocating participants, EE6, showedimpressive performance in both of the conditions. An impor-tant consideration is the fact that EE6 used a finger snap, asopposed to the tongue-click signal used by the otherecholocating participants. As we mentioned in the Methodsection, any participants who used a signal other than tongueclicks or other vocalizations had to keep the source of thesignal (in this case, the hand) close to the body, under the chin,and as still as possible. It is possible, however, that slightmovements were made that were not noticed by the research-er, or possibly even by the participant herself, and that thesemight have aided performance. One might also consider thatthe choice of signal per se could have aided performance; thatis, EE6 used a finger snap, whereas the other EEs used tongueclicks. Yet, several of the control participants used fingersnaps as well, without the advantage that we saw in EE6.

Atten Percept Psychophys

Author's personal copy

In addition to being potentially relevant for explainingEE6’s impressive performance, the question of the choice ofsignal is also relevant for the present study, because themajority of control participants used a signal that was differentfrom those of the echolocating participants (with the exceptionof EE6). Nevertheless, it is important to note that, even thoughthe majority of our EE participants used tongue clicks, this isnot to say that they use this type of signal exclusively ineveryday life. In fact, almost all of the echolocators reportedusing claps, finger snaps, and other techniques. So, the varietyof signals used by echolocators in real life—as well as thevarious signals used by participants in the present study—raises an important question: What is the best signal to use forecholocation? This question has been addressed previously(Rojas, Hermosilla, Montero, & Espi, 2009, 2010), but aconsensus is lacking. For example, longer signals (500 ms)may be better than shorter ones because they result in a‘surplus of echo information’ (Schenkman & Nilsson, 2010).Also, it has been suggested that noise signals provide moreand better information than do click signals (Arias & Ramos,1997), though it has also been suggested that in particular thepalatal tongue click is the best signal for echolocation (Rojaset al., 2009). Therefore, it is unclear whether participants’choices of signals in the present study could have directlyaffected performance (regardless of movement), because thereis no clear indication of what is the best echolocation signal.Also, it is important to note that most systematic studies of thesignals used in echolocation (and many studies on echolo-cation in general) have used artificial sounds played by aloudspeaker or through headphones. Therefore, it will beimportant for future research to address the use of self-produced signals in order to better understand the use ofnatural, active echolocation and to maximize the informa-tion content of echoes.

It can be argued that, at a basic level, the ability toecholocate involves some combination of increased echo sen-sitivity (Dufour, Després, & Candas, 2005; Kolarik, Cirstea,& Pardhan, 2013), suppression of the precedence effect(Wallmeier, Geßele, &Wiegrebe, 2013), and, of course, intacthearing (Schenkman & Nilsson, 2010). This is encouraging,because it means that the ability to echolocate is available toall people, blind or sighted. Therefore, we believe that the useof echolocation should be more actively promoted in the blindcommunity, because even if one learns to echolocate only at avery basic level, it would provide another resource for per-ceiving one’s surroundings and gaining further independencein life. In fact, a recent survey has shown that the use ofecholocation by the blind may have real-world advantages(Thaler, 2013). In particular, blind echolocators have highersalaries and greater mobility in unfamiliar places than do blindindividuals who do not echolocate. Of course, other variableslikely mediate these advantages, but even the additional infor-mation that an echolocator possesses about his or her

surroundings—which then aid in obstacle avoidance, naviga-tion, and object perception—is an advantage in itself.

Overall, the results of the present experiments show thatactive echolocation is a useful resource that allows blindindividuals to gather accurate object shape information fromfaint echoes. Even this basic application of echolocationshows how useful it can be, by providing blind individualswith perceptual information that they would otherwise nothave access to. Considering that echolocation is a trainableskill, it has great potential to offer valuable and liberatingopportunities for the blind and visually impaired.

Author note This work was supported by a Natural Sciences andEngineering Research Council of Canada (NSERC) Scholarship toJ.L.M., and by an NSERC Discovery grant to M.A.G. (Grant No.6313). The authors declare no conflict of interest.

References

Ammons, C. H.,Worchel, P., & Dallenbach, K.M. (1953). “Facial vision”:The perception of obstacles out of doors by blindfolded andblindfolded deafened subjects. American Journal of Psychology, 66,519–553.

Arias, C., & Ramos, O. A. (1997). Psychoacoustic tests for the study ofhuman echolocation ability. Applied Acoustics, 51, 399–419.

Arnott, S. R., Thaler, L., Milne, J. L., Kish, D., & Goodale, M. A. (2013).Shape-specific activation of occipital cortex in an early blind echo-location expert. Neuropsychologia, 51, 938–949. doi:10.1016/j.neuropsychologia.2013.01.024

Cotzin, M., & Dallenbach, K. M. (1950). “Facial vision”: The role ofpitch and loudness in the perception of obstacles by the blind.American Journal of Psychology, 63, 485–515.

Crawford, J. R., & Garthwaite, P. H. (2002). Investigation of the singlecase in neuropsychology: Confidence limits on the abnormalityof test scores and test score differences. Neuropsychologia,40, 1196–1208.

Crawford, J. R., & Garthwaite, P. H. (2007). Comparison of a single caseto a control or normative sample in neuropsychology: Developmentof a Bayesian approach. Cognitive Neuropsychology, 24, 343–372.doi:10.1080/02643290701290146

Crawford, J. R., & Garthwaite, P. H. (2012). Single-case research inneuropsychology: A comparison of five forms of t-test for compar-ing a case to controls. Cortex, 48, 1009–1016. doi:10.1016/j.cortex.2011.06.021

Crawford, J. R., & Howell, D. C. (1998). Comparing an individ-ual’s test score against norms derived from small samples.Clinical Neuropsychologist, 12, 482–486.

Crawford, J. R., Garthwaite, P. H., & Porter, S. (2010). Point and intervalestimates of effect sizes for the case–controls design in neuropsy-chology: Rationale, methods, implementations, and proposedreporting standards. Cognitive Neuropsychology, 27, 245–260. doi:10.1080/02643294.2010.513967

Dufour, A., Després, O., & Candas, V. (2005). Enhanced sensitivity toecho cues in blind subjects. Experimental Brain Research, 165,515–519.

Hardiess, G., Gillner, S., & Mallot, H. A. (2008). Head and eye move-ments and the role of memory limitations in a visual search para-digm. Journal of Vision, 8(1), 7. doi:10.1167/8.1.7. 1–13.

Atten Percept Psychophys

Author's personal copy

Harley, H. E., Putman, E. A., & Roitblat, H. L. (2003). Bottlenosedolphins perceive object features through echolocation. Nature,424, 667–669.

Hausfeld, S., Power, R. P., Gorta, A., &Harris, P. (1982). Echo perceptionof shape and texture by sighted subjects. Perceptual and MotorSkills, 55, 623–632.

Kellogg, W. N. (1962). Sonar system of the blind. Science, 137, 399–404.Kolarik, A. J., Cirstea, S., & Pardhan, S. (2013). Evidence for enhanced

discrimination of virtual auditory distance among blind listenersusing level and direct-to-reverberant cues. Experimental BrainResearch, 224, 623–633. doi:10.1007/s00221-012-3340-0

Niemeier, M., Crawford, J. D., & Tweed, D. B. (2003). Optimaltranssaccadic integration explains distorted spatial perception.Nature, 422, 76–80.

Prime, S. L., Vesia, M. V., & Crawford, J. D. (2011). Cortical mechanismsfor trans-saccadic memory and integration of multiple object fea-tures. Philosophical Transactions of the Royal Society B, 355, 540–553.

Rayner, K. (1998). Eye movements in reading and information process-ing: 20 years of research.Psychological Bulletin, 124, 372–422. doi:10.1037/0033-2909.124.3.372

Rice, C. E. (1967). Human echo perception. Science, 155, 656–664.Rice, C. E. (1969). Perceptual enhancement in the early blind?

Psychological Record, 19, 1–14.Rice, C. E., & Feinstein, S. H. (1965). Sonar system of the blind: Size

discrimination. Science, 148, 1107–1108.Rice, C. E., Feinstein, S. H., & Schusterman, R. J. (1965). Echo-detection

ability of the blind: Size and distance factors. Journal ofExperimental Psychology, 70, 246–251.

Rojas, J. A.M., Hermosilla, J. A., Montero, R. S., & Espi, P. L. L. (2009).Physical analysis of several organic signals for human echolocation:Oral vacuum pulses. Acta Acustica United With Acustica, 95,325–330.

Rojas, J. A.M., Hermosilla, J. A., Montero, R. S., & Espi, P. L. L. (2010).Physical analysis of several organic signals for human echolocation:Hand and finger produced pulses. Acta Acustica United WithAcustica, 96, 1069–1077.

Schenkman, B. N., & Nilsson, M. E. (2010). Human echolocation: Blindand sighted persons’ ability to detect sounds recorded in the pres-ence of a reflecting object. Perception, 39, 483–501.

Schnitzler, H., & Kalko, E. K. V. (2001). Echolocation by insect-eatingbats. BioScience, 51, 557–569.

Schörnich, S., Wallmeier, L., Gessele, N., Nagy, A., Schranner, M., Kish,D., & Wiegrebe, L. (2013). Psychophysics of human echolocation.Advances in Experimental Medicine and Biology, 787, 311–319.

Seibert, A., Koblitz, J. C., Denzinger, A., & Schnitzler, H. (2013).Scanning behavior in echolocating common pipistrelle bats(Pipistrellus pipistrellus). PLoS ONE, 8, e60752. doi:10.1371/journal.pone.0060752

Simmons, J. A., & Chen, L. (1989). The acoustic basis for target discrim-ination by FM echolocating bats. Journal of the Acoustical Societyof America, 86, 1333–1350.

Stoffregen, T. A., & Pittenger, J. B. (1995). Human echolocation as abasic form of perception and action. Ecological Psychology, 7, 181–216.

Supa, M., Cotzin, M., & Dallenbach, K. M. (1944). “Facial vision”: Theperception of obstacles by the blind. American Journal ofPsychology, 57, 133–183.

Teng, S., Puri, A., &Whitney, D. (2012). Ultrafine spatial acuity of blindexpert echolocators. Experimental Brain Research, 216, 483–488.

Teng, S., & Whitney, D. (2011). The acuity of echolocation: Spatialresolution in sighted persons compared to the performance of anexpert who is blind. Journal of Visual Impairment and Blindness,105, 20–32.

Thaler, L. (2013). Echolocation may have real-life advantages for blindpeople: An analysis of survey data. Frontiers in Psychology, 4, 98.doi:10.3389/fphys.2013.00098

Thaler, L., Arnott, S. R., & Goodale, M. A. (2011). Neural correlates ofnatural human echolocation in early and late blind echolocationexperts. PLoS ONE, 6, e20162. doi:10.1371/journal.pone.0020162

Thaler, L., Milne, J. L., Arnott, S. R., Kish, D., & Goodale, M. A. (2013).Neural correlates of motion processing through echolocation,source hearing, and vision in blind echolocation experts andsighted echolocation novices. Journal of Neurophysiology.doi:10.1152/jn.00501.2013

Thomas, J. A., Moss, C. F., & Vater, M. (2004). Echolocation in bats anddolphins. Chicago: University of Chicago Press.

Wallmeier, L., Geßele, N., & Wiegrebe, L. (2013). Echolocation versusecho suppression in humans. Proceedings of the Royal Society B,280, 20131428. doi:10.1098/rspb.2013.1428

Atten Percept Psychophys

Author's personal copy