All testing procedures were approved by the Ethics Board (ECSW2013-1811-150) of the Faculty of Social Sciences at the Radboud University. Participants gave written informed consent prior to testing.
Twenty-three sighted participants were recruited for this study. Participants were undergraduate students at Radboud University and were compensated with course credits. Of the 23 participants, 5 participants did not move their head at all during the echolocation task. As shown in previous research, echolocation experts who were not allowed to move their head during echolocation performed at chance level when determining the shape of an object (Milne et al. 2014b). In our experiment, performance of the five participants that did not move their head did not differ from chance level. This shows that head movements are essential for proper echolocation. Therefore, these five participants were excluded from further analyses, yielding the results of 18 participants (four males), aged between 18 and 24 years (M = 19.50; SD = 1.72). These 18 participants moved their head horizontally, vertically, or both, during trials. We performed pure-tone hearing audiometry in the 250–4000 Hz range to establish that all participants had normal hearing (<20 dB hearing loss).
Stimuli and procedure
The experiment took place on four separate days. On day one and two, participants completed the echolocation task, consisting of two sessions per day. The echolocation task took between 1 and 2 h per day to complete. On day three, participants completed the spatial abilities task, digit span, and the Paced Auditory Serial Attention Task (PASAT). These were completed between 1 and 2 h. On day four, participants’ hearing ability was assessed.
The experiment took place inside a soundproof room (approximately 750 × 480 cm). The ambient sound level in the room was 28 dBA, as measured with a microphone (Bruel & Kjaer, Type 4192) and an amplifier (Bruel & Kjaer, Type 2610). The apparatus that was used to assess participants’ echolocation ability was a size-discrimination task which was based on the apparatus used by Teng and Whitney (2011) and Thaler et al. (2014b). This apparatus is illustrated in Fig. 1. The apparatus consisted of a vertical metal frame with two horizontal metal bars. These bars were spaced 27.5 cm apart. Flat, circular discs that were made from acrylic (0.5 cm thick) and painted with primer could be placed at the end of these bars. The diameter of the largest disc (the reference disc) was 25.4 cm. The diameters of the five comparison discs were 5.1, 9, 13.5, 17.5, and 22.9 cm, respectively. Angular size differences were approximately 22.1°, 18.2°, 13.4°, 9°, and 2.9°, respectively. In total, each participant completed a total of 200 trials, 40 for each comparison disc. The placement of discs was randomized such that each comparison disc appeared equally often on the top and bottom location. A small speaker (JBL Micro II, 150–20.000 Hz) was attached to the forehead of the participant. The speaker was placed in the middle of the forehead with the lower part of the speaker just above the eyebrows. The speaker generated a 10 ms white noise signal (80 dB). We used a speaker to ensure the consistency of the sound signal that was being used. When the participant pressed a button, the sound signal was emitted. The experiment was performed using Presentation software (Version 17.1).
Participants performed the echolocation task in line with the procedure of Teng and Whitney (2011) and Thaler et al.(2014b). Participants were blindfolded and sat at a distance of 320 cm from the end wall of the room. The frame was placed at a distance of 50 cm measured from the discs to the participant’s ear. We opted for a larger distance than the distance (33 cm) used by Teng and Whitney (2011) and Thaler et al. (2014a), because our pilot studies showed a ceiling effect in performance at 33 cm. As a consequence, performance was high for all participants, showing almost no variance within the group, and excluding learning to occur. The height of the frame was adjusted in such a way that the participant’s ear was halfway between the two bars. The speaker was attached to the participants’ forehead (centre), and the button was placed in the hand of the participant. The participant completed two practice trials, which were followed by 50 test trials. After a short break of 5 min, another 50 trials followed. Between the trials, participants pressed their ears with their fingers to prevent them from hearing the experimenter placing the reference disc (i.e. the larger disc) on either the top or bottom bar and one of the five comparison discs on the remaining bar. In addition, even if a trial contained the same disc set-up as the trial before, the experimenter took the discs of the bar, and placed them back. This way, possible hints from noise from placing the discs or the amount of time it took the experimenter to place the discs, would be uninformative. Once the discs were placed, the experimenter tapped the participant on the shoulder to indicate that they could unblock their ears. First, the participant was asked to indicate whether they thought that the reference disc was on the top or bottom bar, with refraining to use sound (‘no-sound judgment’). By obtaining a no-sound judgment, a level of baseline performance or possible effects of background noise were established. This no-sound judgment might take into account any information from ambient noise or noise that arose from placing the discs. After the no-sound judgment, participants pressed the button to emit the sound signal from the speaker. Participants were allowed to produce sounds for up to 20 s to determine whether the reference disc was on the top or bottom bar (‘sound judgment’). The amount of emitted sound signals per trial and the time spent to make a judgment were registered. After each sound judgment, we registered the level of confidence of the participant concerning that sound judgment. Participants indicated how certain they were of their judgment on a scale of 1 (very uncertain) to 5 (very certain). Because we are interested in learning of echolocation, participants received correct/incorrect feedback after every trial (Herzog and Fahle 1997), in contrast to Teng and Whitney (2011) and Thaler et al. (2014a).
To measure participants’ spatial abilities in a non-visual manner, we used the following set-up, which was an adaptation of the set-up used by Simons and Wang (1998) and Pasqualotto and Newell (2007). A rotatable circular platform (70 cm in diameter) which contained 36 sunken position markers was placed on a table. In these markers, 6 objects (rectangular solid, cube, cylinder, sphere, cone, and pyramid) could be placed (see Fig. 2). The objects were placed on the platform at a random location, with the only constraint that the objects would have a minimal distance of 10 cm to each other. Subjects were blindfolded and sat in front of a table on which the platform was placed. In each trial, participants were given 60 s to learn where the objects were placed via haptic exploration. Subsequently, the experimenter randomly removed one or more objects from the platform and turned the platform 90° clockwise. Next, the participant placed the object(s) back on the (rotated) platform without time restriction. The number of objects that was removed from the platform increased every two trials, starting with one object in the first two trials, then two objects in the next two trials, etcetera. The possible total number of test trials was twelve. Participants completed three practice trials with feedback, followed by the test trials without feedback. The object was judged as correct if it was placed in the exact location or the location adjacent to that one. The test ended if the subject failed placing all objects correctly on the platform in two consecutive test trials with the same number of objects removed. The total score consisted of the number of correct trials with the maximum possible score of 12.
To assess working memory, we used the digit span subtest of the Dutch Wechsler Adult Intelligence Scale (WAIS-IV-NL; Wechsler 2012). In this task, participants were asked to repeat a string of digits in the same order (forward recall), reversed order (backward recall), or ascending order (sort recall). These digits were presented verbally by the experimenter. The number of digits in a string increased every two trials. The test ended if the participant incorrectly repeated the string of digits with the same length twice. The total score consisted of the total number of correct series for the three tests combined. The total score was converted into a standard score based on Dutch norms. This task has good test–retest reliability, with a correlation coefficient of .88 (Wechsler 2012).
Sustained and divided attention
Sustained and divided attentions were measured using the Dutch version of the Paced Auditory Serial Addition Task (PASAT; Aarnoudse et al. 1995; Gronwall 1977). In this task, 60 single-digit numbers were presented to the participant. Participants were asked to add each digit to the one immediately preceding it and to immediately give the sum of these two digits. The test consisted of five blocks of 60 numbers with decreasing intervals between digits, ranging from 3.2 to 1.6 s. One total score was calculated, which consisted of the total number of correct additions, calculated over five blocks. The test–retest reliability of this task is good, with a correlation coefficient of .82 (Aarnoudse et al. 1995).
The dependent variable was the percentage of correct judgments, which was calculated separately for each of the four sessions (one, two, three, and four), signal (no-sound and sound), and for each of the five angular size differences between the reference disc and the comparison disc (2.9°, 9°, 13.4°, 18.2°, and 22.1°). Chance performance was 50%. A three-way repeated measures ANOVA with within-subject factors session, signal, and angular size difference was computed. If the variances were not homogeneous (as determined with Mauchly’s test), the Greenhouse-Geisser correction was used. Furthermore, correlation analysis was conducted using correlation coefficients to determine the relationship between, on the one hand, the increase in echolocation ability and, on the other hand, the cognitive measures. We calculated the increase in echolocation ability as the difference between the sound and no-sound performance, looking at the difference in performance between the last and first session, across the angular size differences. An alpha level of .05 was considered statistically significant.