D

Journal of Communication and Computer 9 (2012) 560-567

DAVID

PUBLISHING

Presence, Physiological Arousal, and Visual Recognition in 3D TV Soyoung Bae, Chris Eller and Annie Lang Department of Telecommunications, Indiana University, Bloomington, IN 47405, USA

Received: July 25, 2011 / Accepted: August 30, 2011 / Published: May 31, 2012. Abstract: Stereoscopic (3D) presentations have become increasingly popular and common in both movie theatres and in everyday devices, such as televisions and cameras. But why do they appeal to viewers? To answer this question, it is necessary to assess whether 3D presentation provides any benefits to viewers, either for entertainment or cognitive processing, compared to monoscopic (2D) presentation. Thus, this study examined the effect of dimensionality on presence, arousal, and visual recognition. The results showed that film clips in 3D evoked more presence than 2D clips did. However, compared to 2D, 3D did not necessarily elicit higher arousal or visual recognition. Key words: 3D TV, presence, physiological arousal, visual recognition.

1. Introduction There have been many attempts to develop stereoscopic displays in TVs, movie theatres, computers, and other electronic devices (e.g., cameras). Currently, most people experience 3D movies in movie theatres. Broadcasting companies have also begun to generate 3D content and 3D TVs are becoming increasingly popular. To implement these technologies researchers have analyzed audience perception of 3D videos (See [1]) and technical problems related to the generation and delivery of stereoscopic television (See [2]). As a result we are now capable of delivering 3D content to viewers’ televisions and we know that viewers find 3D content entertaining. What we do not know how the addition of 3D alters message processing and viewers’ emotional and cognitive responses is. Except in the field of engineering, most studies of 3D have been done in the field of human-computer interaction and tend to examine whether a 3D Corresponding author: Soyoung Bae, Ph. D. candidate, research fields: embodied, embedded, dynamic motivated cognition, Human Media Interaction. E-mail: [email protected].

presentation helps with computer task performance and enjoyment. For example, a series of studies of computer interface systems have compared 2D and 3D computer interfaces on spatial memory [3-6]. However, little research has been done investigating whether 3D videos elicit different experiences (e.g., arousal, presence, etc.). Furthermore, we do not know how 3D videos influence information processing. While 3D videos may be more arousing and enjoyable; they may also elicit greater fatigue than 2D videos. Since most people watch to relax a more fatiguing format might be counterproductive. This study is designed to whether 3D compared to 2D content elicits more presence and arousal in viewers and requires more cognitive resources to process.

2. Literature Review First, it is expected that 3D content will increase viewers’ sense of presence. Presence is a psychological construct and has been defined in a variety of ways. For some presence includes feeling involved with the subject matter [7]. Others have defined presence as having multiple dimensions

Presence, Physiological Arousal, and Visual Recognition in 3D TV

including realism, transportation, and immersion. The conceptualization of presence as realism emphasizes “the degree to which a medium can produce seemingly accurate representations of objects, events, and people–representations that look, sound, and/or feel like real thing” [8].When conceptualized in terms of transportation, presence is defined as an awareness that either the viewer is in the space of the film or the object is here in proximity to the viewer. The viewer may feel as though they are transported to another place or that another place is transported to the viewer. Presence as immersion refers to perceptual and psychological immersion or “the degree to which a virtual environment submerges the perceptual system of the user” [9]. Not only virtual reality systems but also IMAX theaters and even traditional movie theaters can be included in this definition because they immerse the senses of media users. The width and height of the screen are regarded as the X-axis and the Y-axis, respectively. A perceived depth of 3D movies can be described as a Z-axis, which 2D movies lack. When one watches 3D stereoscopic movies, they give the illusion of looking at different layers or distances on a screen. In other words, some objects appear to be in front of the screen while other objects appear behind the screen. Therefore, 3D movies allow viewers to experience movie characters and elements that appear to be in a real space (i.e., realism). Moreover, viewers may experience a sense of “being there” (transportation) and viewers may become more involved and absorbed (immersion) than they do when watching 2D movies. In addition, presence is defined as how one perceives one’s environment through the use of mental processes, some of which are consciously controlled and others that are automatic in nature [10]. A stimulus from multiple sensory channels, an increase in information bandwidth, or the degree of depth of a single stimulus results in a greater sense of presence when compared to a lower bandwidth, or a

561

shallower version of the same stimulus [11]. This study was able to deliver stereoscopic content via the visual sense, thus affecting presence as experienced through the visual sensory channel. Dimensionality as one of the characteristics of visual display could cause presence [8]. Given these definitions it seems likely that 3D presentation will add to realism, immersion, and information bandwidth. Therefore, this study predicts: H1: 3D clips will evoke a stronger sense of presence than 2D clips. Presence is often related to arousal [12]. Arousal has been conceived of as a drive state or as the intensity of an emotion or experience [13-14]. Movies, virtual reality entertainment systems, or video games that elicit presence are usually designed to be arousing [8]. A study [15] found that when compared to 2D, 3D evoked greater physiological arousal while people were watching sports and commercials on television. In addition, 3D made viewers experience stronger presence, enjoyment, and involvement. Likewise, Freeman and Avons [16] found that people reported more involving and intensifying experiences during 3D compared to 2D excerpts. On the other hand, some studies have found little support for the hypothesis that arousal is one of the physiological effects of presence [17-18] while others suggest that the relationship between presence and arousal is context dependent (e.g., [19-20]). For example, Rajae-Joordens [21] tested the effect of dimensionality on presence, physiological arousal and recognition memory while subjects either played a game or watched a movie. When playing a game in 3D, people felt greater presence and physiological arousal when compared to the same game in 2D. Watching a movie in 3D did not result in higher presence or arousal when compared to the same movie in 2D. Although this finding might be problematic because it used only one stimulus per dimensionality condition, we should exercise caution about

562

Presence, Physiological Arousal, and Visual Recognition in 3D TV

generalizing the relationship between presence and physiological arousal. Thus, a research question was posed: RQ 1: Will 3D clips elicit more arousal than 2D clips? In addition to the research on the relationship among 3D, presence, and arousal, research has also examined the effect of 3D on memory. In the field of education, Bronack and his colleagues tested how 3D virtual worlds contributed to the social aspects of learning and teaching as compared to traditional circumstances [22]. However, this study did not compare the difference between 3D and 2D. In the area of human computer interface; several studies have investigated how the 3D interface has influenced spatial memory in a computer document or a window interface. One study showed that recall of the locations of the alphabet was more effective when using 3D interface [6]. Another study examined whether a 3D window manager provided better support for document comparison and task management. Usability tests showed that these3D prototypes were enjoyable and evoked good spatial memory. Additionally, people showed a stronger sense of recalling front-to-back orders in tasks compared to recalling left-to-right orders in 3D space [3].Not only in real environment but also in virtual environment, 3D showed an impact on cognition. Cockburn and McKenzie [4] investigated the effectiveness of spatial memory in 2D and 3D physical and virtual environments for graphic computer interfaces. For both physical and virtual environments, users had greater difficulty in retrieving information from 3D interfaces than they did from 2D interfaces. Similarly, Westerman and Cribbin [23] found that in virtual reality, 3D representation caused higher cognitive demand compared to 2D representation. In line with the findings of the studies done on computing environments, some studies have focused on TV or 3D movies. Watching 3D commercials on

TV led higher recognition and cued recall compared to watching

2D

commercials

[15].

In

contrast,

Rajae-Joordens [21] showed that watching a movie in 3D did not result in higher recognition memory compared to 2D. There are several reasons why these memory results are inconsistent. Most importantly, 3D in different media or environments might require differing levels of cognitive resources to process information. Additionally, some tasks in those environments differ from comparable tasks in 2D environments. For example, computer interface systems require users to operate them and remember document placement. Similarly, in virtual realities, people need to actively participate in order to browse, operate, or manipulate that environment. In contrast, traditional 2D TV does not require much active participation from its users. Therefore, using 3D presentations in TV applications may require us to consider different patterns of information processing. Another reason for the inconsistent memory results could be that these studies have used different memory measures: recognition, cued recall, and recall. These measures are recognized as indicators of different sub-processes of memory [24]: encoding, storage, and retrieval, respectively. Recognition is regarded as an indicator of encoding specific information. For example, multiple choice questions and yes/no questions are used for recognition. Cued recall is indexed as a reflection of how well encoded information is stored. A cue is presented to help a person to retrieve information. Retrieval is an index of how well stored information can be retrieved without any cues. In this study, we are not interested in storing or retrieving information. Instead, this study focuses on encoding information from short movie clips. Since this study uses only video without audio, only visual recognition will be measured. Based on the aforementioned discussion, a research question is posed: RQ2: Will visual recognition for 3D clips be higher than visual recognition for 2D clips?

Presence, Physiological Arousal, and Visual Recognition in 3D TV

3. Method 3.1 Stimuli Stereoscopic and monoscopic versions of everyday scenes on a large college campus were recorded and edited to create pairs of stimuli where the monoscopic version used the footage from one camera while the stereoscopic version used the footage from both cameras. A total of 12 movieclips were used as stimuli. The clips ranged in duration from 30-60 seconds. The clips represented a variety of settings, subject matters, objects and camera movements. The clips included an artist painting outdoors, a marching band performing at a football game, a play performed on stage, a dance group performance, people playing Frisbee in a grassy area, an opera being performed on stage, to name a few(see Fig. 1). The clips varied enormously in amount and complexity of audio information. Therefore, in order to control audio information level, all audio was removed when the stimuli were presented to subjects. These clips were divided into three groups of similar duration (about 3 minutes and 30 seconds). Subjects were randomly assigned to group. The clips were generated from footage gathered by an undergraduate video production class, a graduate video production class and professional staff at the same university. The video clips were created using two different stereoscopic camera rigs:  Sony XDCAM EX3 cameras at full 1920 × 1080 resolution: 1:1 pixel aspect ratio; In a state of gunlock.  Sony Z7U HDV cameras at full 1440 × 1080 resolution: 1.21:1 pixel aspect ratio; Synchronized with a LANC Shepherd sync device.

563

The number of sensory channels, which mainly distinguish between the stereoscopic and monoscopic content [10], was engaged. The monoscopic content indicates that a single stream of content was delivered identically to both eyes at the same time. The stereoscopic content means that two streams of data are each delivered to only the appropriate eye. In an ideal situation, the data delivered would be doubled as compared to a monoscopic version of the media; however, the display used for the study displayed half-resolution per eye. The sum of the pixels displayed was still the same, but the pixels shown per eye were from different data streams as generated by the two cameras in the stereoscopic video rig. The stereo monitor uses a 3D display technology called Xpol, in which alternating lines of the display are used to show the left and right images in a horizontal interleaved format. Each line is then polarized with a micropolarizer that corresponds to the polarized lenses in the viewer’s glasses: the odd-numbered lines were polarized counterclockwise and the even-numbered lines were polarized clockwise. Each eye was then presented with a display that had full horizontal resolution, but half vertical resolution. 3.2 Design This study featured a Dimensionality (2) × Repetition

(2)

×

Group

(3)

mixed

design.

Dimensionality and Repetition were within-subjects factors while Group was a between-subjects factor. Each subject watched four clips: two in 2D and the other two in 3D. For example, Subject A watched clips 01 and 03 monoscopically and clips 04 and 08 stereoscopically. Another subject, Subject B, watched clips 01 and 03 stereoscopically and clips 04 and 08 monoscopically. The order in which these stimuli were presented was randomized. 3.3 Participants

Fig. 1

Sample screen shots of stimuli.

Twenty-six participants (Males = 16, Females = 10) from classes within a communications department at a

564

Presence, Physiological Arousal, and Visual Recognition in 3D TV

large Midwestern university participated in this study for extra credit. When answering questions about their binocular vision, none of the participants reported a disability. In addition, all subjects reported that they had seen at least a couple of 3D movies previously. 3.4 Dependent Variables 3.4.1 Presence Items from the Temple Presence Inventory (TPI) were used to measure presence. The spatial presence, social presence, engagement, and perceptual realism sub-categories of the TPI were used in a self-report questionnaire. The stimuli did not provide sound or touch-based experience, so those questions were excluded. The stimuli did not have a significant amount of narrative, so parasocial-related questions were also excluded. Questions on the inventory included “How much did it seem as if the objects and people you saw had come to the place you were?” “How often when an object seemed to be headed toward you did you want to move to get out of its way?” “Did the experience seem more like looking at the events/people on a movie screen or more like looking at the events/people through a window?” and so on. The scale of the responses ranged from 1 (e.g., not at all, never, like a movie screen, etc.) to 7 (e.g., very much, always, like a window, etc.). These ratings were averaged to create an index from 1 meaning little presence to 7 meaning a great deal of presence. 3.4.2 Physiological Arousal Physiological arousal was measured by skin conductance. Skin conductance has been recognized as an operational definition of physiological arousal [25-26]. The experiment was controlled by a 386-Mhz computer running VPM experimental software with a LabMaster AD/DA board installed. Skin conductance data were collected using a Coulbourn S-71-22 coupler sampling conductance levels 20 times per second from two standardized (8 mm recording surface) AG/AGCL electrodes filled with non-conductive gel. Electrodes were attached to the

palmar surface of the subject’s non-dominant hand, which was cleaned and hydrated with distilled water. Signals were aggregated for analysis as the highest level per second. 3.4.3 Recognition Memory A speeded forced choice visual recognition test was used to index encoding. Scenes from the portions of the clips that subjects had seen were used as targets, while images from portions of the clips that the subjects did not see were used as foils. Subjects pushed a yes or no key to indicate if they had seen the scene previously. The scenes were presented for 75 ms and both accuracy and latency of response were recorded using Media Lab (Empirisoft). All images were presented monoscopically due to hardware and software limitations. 3.5 Procedure Upon arrival, the participants were greeted and introduced to the experiment. Participants were tested individually. A pair of 3D glasses (circular polarized) was given to the participant to watch a 3D calibration clip, which was needed to adjust the angle of the stereoscopic display relative to the subject’s head. The stereoscopic display’s effectiveness was dependent on the vertical angle of view. While the subject was calibrating his or her view on the calibration clip, two electrodes for skin conductance were attached to the palm of the participant’s non-dominant hand. After the calibration was completed, the clips were shown. After watching each clip, the subject completed the self-reported presence questionnaire. When the subject had finished watching all four movies (two in 2D and two in 3D), the electrodes were removed. Participants viewed a short film to clear short term memory, completed the visual recognition test, and were thanked and dismissed.

4. Results A repeated measures analysis of variance (ANOVA), Dimensionality (2) × Repetition (2) ×

Presence, Physiological Arousal, and Visual Recognition in 3D TV

Group (3), was conducted for analysis. The hypothesis that 3D clips (M = 3.549, SD = 0.125) would evoke greater presence than 2D clips would (M = 2.647, SD = 0.116) was supported, F(1, 23) = 34.030, p < 0.001. That is, 3D clips did provide more presence than 2D. Research question 1 asked if 3D clips would elicit more physiological arousal than 2D clips. The data for two subjects were lost due to technical problems, resulting in an n of 24. The results were not significant (F < 1). Therefore, there was no difference in physiological arousal when the subjects were watching 2D or 3D clips. The second research question asked whether visual recognition would be better for 3D compared to 2D clips. The same ANOVA was run on the visual recognition data. Again, the result was not significant (F < 1). Thus, the visuals from 3D were not encoded any better than those of 2D.

5. Discussion This study investigated whether 3D versions of film clips would elicit greater presence, arousal, and visual memory compared to 2D versions of the same clips. Results showed a significant effect of 3D on presence but no measurable differences in physiological arousal or recognition. According to this study, it appears that while 3D videos increase presence, that perception is not associated with increased arousal or better memory. However, there should be some consideration of the limits of this study before these findings are generalized. One major limitation of the study relates to the stimuli used. Stimuli were created for this study in order to be certain that the only difference between conditions was dimensionality. However, this means that the stimuli were relatively simple clips. They did not have strong narratives or much in the way of structural features. As a result, the clips were not particularly involving. Differences between 3D and 2D on memory and arousal might occur if the clips themselves were more exciting.

565

Similarly, it might be that differences between 3D and 2D are not global but are local. For example, it may be that certain structural features (such movement towards and away) or production features (e.g., scene changes, zooms, etc.) might have more impact on arousal and memory in 3D compared to 2D versions. Future research should investigate the local impact of 3D on arousal and memory. In addition, these clips had no audio track which probably reduced realism and transportation. Future research should assess whether processing audio information is improved or damaged by the presence of 3D content. Another possible limitation lies in our choice of monitor. It is possible that the display size of 22 inches (at a resolution of 1680 × 1050 total, with a resolution of 1680 × 525 per eye in stereo) may not have been large enough to optimize the perception of3Ddepth cues to distinguish the 3D visual information from the 2D information. Finally, it may be that presenting the 3D recognition test in 2D rendered the task harder for the 3D clips compared to the 2D clips since the scenes had not been seen in 2D and thereby masked any effect of dimensionality on recognition. Future research should not only consider these limitations but also investigate possible individual differences. This study had a small sample size which made sub-group investigation impossible. For example, we know that there are significant gender differences in spatial cognition. Overall, men outperform women on most spatial tasks, such as mental rotation and route-finding [27-28]. However, some studies have shown that women outperform men in object location memory [29-33]. It would be interesting to determine if gender plays a role in the processing and experience of 3D verses 2D content. Finally, it would be interesting to investigate whether previous experience with 3D makes any difference in people’s current experiences with 3D. As of this writing, 3D has just begun to be

Presence, Physiological Arousal, and Visual Recognition in 3D TV

566

commercialized. Therefore, people are not yet very used to 3D. That is one reason why watching 3D could be more entertaining. More exposure to 3D might decrease the effect of dimensionality because people will become used to it. For example, arousal and presence for the first-time user might be higher than those of users who have had more experience.

[8]

[9]

[10]

6. Conclusions These results show that people experienced higher presence during 3D clips but showed no influence of 3D on physiological arousal or recognition memory. Future research on 3D should investigate questions related to structural features, content, modality, gender and media use.

[11] [12]

[13]

References [1]

[2]

[3]

[4]

[5]

[6]

[7]

S.L.P. Yasakethu, W.A.C. Fernando, B. Kamolrat, A. Kondoz, Analyzing Perceptual Attributes of 3D Video, IEEE Transactions on Consumer Electronics 55 (2009) 864-872. S.J.E. Jolly, M. Armstrong, R.A. Salmon, Three-dimensional television; a broadcaster’s perspective, in: Proceedings of SPIE, 2009, pp. 72370B-72371-77237B-72311. G. Robertson, M. Van Dantzich, D. Robbins, M. Czerwinski, K. Hinckley, K. Risden, et al., The task gallery: A 3D window manager, in: Proceedings of CHI'2000 Conference on Human Factors in Computing Systems, The Hague, Netherlands, 2000, April. A. Cockburn, B. McKenzie, Evaluating the effectiveness of spatial memory in 2D and 3D physical and virtual environments, in: CHI ’02: Proceedings of the SIGCHI conference on Human factors in computing systems, ACM Press, New York, NY, 2002, pp. 203-210. A. Cockburn, B. McKenzie, Evaluating spatial memory in two and three dimensions, International Journal of Human-Computer Studies 61 (2004) 359-373. M. Tavanti, M. Lind, 2D vs. 3D, implications on spatial memory, in: Proceedings of IEEE Info Vis 2001 Symposium on Information Visualization, San Diego, CA, 2001, October. M.T. Palmer, Interpersonal communication and virtual reality: Mediating interpersonal relationships, in: F. Biocca, M.R. Levy (Eds.) Communication in the age of virtual reality, Lawrence Erlbaum, Hillsdale, NJ, 1995, pp. 277-302.

[14]

[15]

[16]

[17]

[18]

[19] [20]

M. Lombard, T. Ditton, At the heart of it all: The concept of presence, Journal of Computer Mediated Communication 3 (1997) 1-40. F. Biocca, B. Delaney, Immersive virtual reality technology, in: F. Biocca, M.R. Levy (Eds.) Communication in the age of virtual reality, Lawrence Erlbaum Associates, Hillsdale, NJ, 1995, pp. 57-124. J. Steuer, Defining virtual reality: Dimensions determining telepresence, in: F. Biocca, M.R. Levy (Eds.) Communication in the age of virtual reality, Lawrence Erlbaum Associates, Hillsdale, NJ, 1995, pp. 33-56. B. Reeves, C. Nass, Perceptual bandwidth, Communications of the ACM, 2000, pp. 65-70. M. Meehan, B. Insko, M. Whitton, J. Frederick, P. Brooks, Physiological measures of presence in stressful virtual environments, ACM Transactions on Graphics, 2002, pp. 645-652. M. Bradley, Emotional theory: A dimensional analysis, in: S. Van Goozen, N.E. Van de Poll, J.A. Sergeant (Eds.), Emotions: Essays on emotion theory, Erlbaum, Hillside, NJ, 1994, pp. 1-37. G. Mandler, Memory arousal and mood: A theoretical integration, in: S. Christianson (Ed.), The Handbook of emotion and memory: Research and theory, Lawrence Erlbaum Associates, Hillsdale, NJ, 1992. D. Varan, A. Bulgrin, 3D TV: Viewer interaction with the new dimension for television, in: The meeting of Audience Measurement 6.0 Agenda, NYC, NY, 2011, June. J. Freeman, S.E. Avons, Focus groups exploration of presence through advanced broadcast services, in: Proceedings of the SPIE, San Jose, CA, 2000, pp. 3959-3976. C. Heeter, Communication research on consumer VR, in: F. Biocca, M.R. Levy (Eds.), Communication in the age of virtual reality, Lawrence Erlbaum Associates, Hillsdale, NJ, 1995, pp. 191-218. M. Lombard, R.D. Reich, M.E. Grabe, C.M. Campanella, T.B. Ditton, Big TVs, little TVs: The role of screen size in viewer responses to point-of-view movement, in: International Communication Association, Albuquerque, NM., 1995, May. E.L. Salnass, in: The 2nd International Workshop on Presence, University of Essex, Colchester, 1999, April. B.K. Wiederhold, R. Davis, M.D. Wiederhold, Correlation of physiological arousal and immersion levels in virtual worlds, in: G. Riva, B.K. Wiederhold (Eds.), Virtual Environments in Clinical Psychology and Neuroscience: Methods and Techniques in Advanced Patient-Therapist Interaction, IOS Press, Amsterdam 1998, pp. 52-60.

Presence, Physiological Arousal, and Visual Recognition in 3D TV [21] R.J.E. Rajae-Joordens, Measuring Experiences in Gaming and TV Applications, in: J.H.D.M. Westerink, M. Ouwerkerk, T.J.M. Overbeek, W.F. Pasveer, B. Ruyter (Eds.), Probing Experience, Springer Netherlands, 2008, pp. 77-90. [22] Bronack, R. Sanders, A. Cheney, R. Riedl, J. Tashner, N. Matzen, Presence Pedagogy: Teaching and Learning in a 3D Virtual Immersive World, International Journal of Teaching and Learning in Higher Education 20 (2008) 59-69. [23] S.J. Westerman, T. Cribbin, Mapping semantic information in virtual space: dimensions, variance, and individual differences, International Journal of Human-Computer Studies 53 (2000) 765-787. [24] A. Lang, The limited capacity model of mediated message processing, Journal of Communication 501 (2000) 46-70. [25] R. Hopkins, J.E. Fletcher, Electrodermal measurement: Particularly effective for forecasting message influence on sales appeal, in: L. A. (Ed.), Measuring psychological responses to media, Lawerence Erlbaum Associates, Inc., Mahwah, NJ, 1994, pp. 113-132. [26] R.F. Potter, P.D. Bolls, Psycho physiological Measures and Meaning: Cognitive and Emotional Processing of Media, Routledge, New York, NY, 2011.

567

[27] D. Voyer, S. Voyer, M.P. Bryden, Magnitude of sex differences in spatial abilities: A meta-analysis and consideration of critical variables, Psychological Bulletin 117 (1995) 250-270. [28] E. Coluccia, G. Louse, Gender differences in spatial orientation: A review, Journal of Environmental Psychology 24 (2004) 329-340. [29] M. Eals, I. Silverman, The hunter-gatherer theory of spatial sex differences: Proximate factors mediating the female advantage in recall of object arrays, Ethology and Sociobiology 15 (1994) 95-105. [30] T.W. James, D. Kimura, Sex differences in remembering the locations of objects in an array: Locationshifts versus location-exchanges, Evolution and Human Behavior 18 (1997) 155-163. [31] L.J. Levy, R.S. Astur, K.M. Frick, Men and women differ in object memory but not performance of a virtual radial maze, Behavioral Neuroscience 119 (2005) 853-862. [32] D.H. McBurney, S.J.C. Gaulin, T. Devineni, C. Adams, Superior spatial memory of women: Stronger evidence for the gathering hypothesis, Evolution and Human Behavior 18 (1997) 165-174. [33] Q. Rahman, S. Abrahams, F. Jussab, Sex differences in a human analogue of the Radial Arm Maze: the “17-Box Maze Test”, Brain Cognition 58 (2005) 312-317.