Asymmetrical Switch Costs in Spatial Reference Frames Switching

Abstract

Previous studies found that the egocentric and allocentric reference frames are distinct in their functions, developmental trajectory, and neural basis. However, these two spatial reference frames exist in parallel, and people switch between them frequently in their daily lives. Using an allocentric and egocentric switching task, this study explored the cognitive processes involved in the switch between egocentric and allocentric reference frames and the possible asymmetry of switch costs. Sixty-two participants were tested in congruent (i.e., the target was on the same side in two reference frames) and incongruent conditions (i.e., the target was on a different side in two reference frames). The results indicated that the interaction between allocentric and egocentric reference frames was bidirectional and that the congruency effect was higher in the egocentric task than in the allocentric task. More important, the switch costs between allocentric and egocentric reference frames were found in both conditions, and the switch cost was higher for allocentric task. To our knowledge, this was the first study to focus on how switch costs and asymmetrical switch costs occur in allocentric and egocentric task switching.

Keywords

task switching allocentric/egocentric reference frames asymmetry of switch costs

Introduction

In daily activities, people need to maintain spatial information to represent the locations of objects around them. The location of an object could be represented in an allocentric (object-to-object) or an egocentric (subject-to-object) reference frame (e.g., Burgess, 2006; Klatzky, 1998; Mou, McNamara, Valiquette, & Rump, 2004). In an allocentric reference frame, the locations are represented within a framework external to the perceiver and independent of his or her position, whereas in an egocentric reference frame, they are represented with respect to the perceiver’s perspective (Klatzky, 1998). For example, when a right-handed person is drinking coffee while using a computer, the mouse is likely to be on his or her right (egocentric representation) but to the left of coffee cup (allocentric representations). The interaction between allocentric and egocentric reference frames and their own features is the focus of studies of spatial cognition.

Two Spatial Reference Frames

The allocentric and egocentric reference frames are distinct from each other in properties and functions. The egocentric reference frame helps to form the transient self-object associations and plays a major role in spatial updating, whereas the allocentric system helps to form the spatial relations among objects and then constructs enduring spatial representations (Burgess, 2006; Mou et al., 2004). The distinction between these two reference frames is also supported by developmental studies. Egocentric spatial coding and updating occurs in the first year (Acredolo, 1978; Landau & Spelke, 1988), while the use of allocentric reference frames emerges in the second year (Nardini, Atkinson, & Burgess, 2008). Furthermore, these two reference frames have distinct developmental trajectories, and the egocentric reference frame remains dominant in early years of life (e.g., Nardini, Burgess, Breckenridge, & Atkinson, 2006).

Neuropsychological findings have shown that egocentric representations are associated with the dorsal stream, while allocentric representations are associated with the ventral stream (Foley, Whitwell, & Goodale, 2015; Goodale & Milner, 1992; Milner & Goodale, 2008). Using functional magnetic resonance imaging, a multitude of studies have found that egocentric tasks are associated with the parietofrontal network and premotor regions, while allocentric tasks are associated with parietotemporal regions (Chen et al., 2014; Galati et al., 2000; Galati, Pelle, Berthoz, & Committeri, 2010; Klatzky, 1998; Neggers, Van der Lubbe, Ramsey, & Postma, 2006; Schenk, 2006; Vallar et al., 1999). Moreover, the hippocampus is related to allocentric tasks (Bohbot, Iaria, & Petrides, 2004; Galati et al., 2000; Parslow et al., 2004). It only receives input from the nonegocentric system, and it does not store egocentric representations (O’Keefe & Nadel, 1978).

Although egocentric and allocentric reference frames are distinct in their functions, developmental trajectory, and neural basis, they exist in parallel (Burgess, 2006). The locations of objects in the environment are not only represented according to one’s own perspective but also stored in representations concerning the external environment. The coexistence of two reference frames raises the question of how they interact. A developmental study demonstrated the parallel operation of these two reference frames and revealed a cooperative effect as early as the age of 3 years (Nardini et al., 2006). Other evidence, in contrast, showed that there might be conflict between the two reference frames. For example, when a background stimulus is placed behind a target object, the judgement of the position of the object relative to the body is biased. Specifically, when the background stimulus is shifted to the left or right, the egocentric position of the target object is perceived to be shifted in the opposite direction (Bridgeman, Peery, & Anand, 1997; Roelofs, 1935). The reverse was also true. The position of a target object relative to the surrounding environment is also influenced by its position relative to the observer’s body (Sterken, Postma, De Haan, & Dingemans, 1999). That is, egocentric and allocentric reference frames could interact with each other.

The allocentric–egocentric task (AET) was developed to investigate the interaction between the two reference frames (Galati et al., 2000; Neggers et al., 2006; Neggers, Scholvinck, van der Lubbe, & Postma, 2005). In a typical AET paradigm, participants are represented with stimuli in which a vertical target bar is placed on a horizontal background bar. In the egocentric task, participants are asked to judge the position of the target bar with respect to the midsagittal plane of their bodies, while in the allocentric task, they are asked to judge the position of the target bar with respect to the centre line of the background bar. Therefore, there are two kinds of stimuli: In the congruent trials, the representations of the target’s position regarding these two reference frames are congruent (egocentric left and allocentric left, or egocentric right and allocentric right). In the incongruent trials, in contrast, they are conflicted regarding these two reference frames (egocentric left and allocentric right, or egocentric right and allocentric left). Previous studies have found that the egocentric task is affected by the incongruent allocentric representations, but not vice versa (Neggers et al., 2005, 2006). Neggers et al. (2005) suggested that the representation of a target in the allocentric reference frame is stronger than in the egocentric one and hence exerts more influence.

In summary, the aforementioned studies suggest that the two spatial reference frames may interact with each other, that their relative strength plays an important role in determining the interaction, and that people usually switch from one reference frame to the other (Waller & Hodgson, 2006). The switch between these two reference frames is worthy of attention, as it happens in many daily spatial activities. Usually, we represent locations egocentrically first and then form an enduring allocentric representation involving many spatial relations based on the egocentric representations. When we decide to walk towards or reach for a target, the action-oriented egocentric representations must be derived from the preceding allocentric representations (Burgess, 2006). Furthermore, the switch in processing between different reference frames may imply the process of encoding, selecting, or combining them to do appropriate actions to meet spatial task requirements (e.g., Ruggiero, Iavarone, & Iachini, 2018). For example, in two recent studies (Ruggiero et al., 2018; Ruggiero, Ruotolo, & Iachini, 2018), participants were invited to memorise the positions of three objects and then to make a spatial judgement about relative distances between memorised stimuli in a switch condition (allocentric to egocentric task or egocentric to allocentric task). Ruggiero et al. found increased switch costs in congenitally blind participants, in patients with Alzheimer’s disease, and in patients with mild cognitive impairments. These increases relative to the performance of healthy participants were observed when the participants were switching from an allocentric to an egocentric reference frame. More important, for all groups, the performance was better when the anchor point was egocentric (i.e., Ego–Ego and Ego–Allo spatial judgement tasks) than when the anchor point was allocentric (i.e., Allo–Allo and Allo–Ego spatial judgement tasks). These authors suggested that this egocentric facilitation was due to the fact that egocentric encoding forms the primary body–environment interface, whereas additional processing would be required to work out allocentric representations (Ruggiero et al., 2018). This finding indicated that encoding egocentric representation would facilitate the processing of allocentric representations, but not vice versa. Recent studies have begun to focus on the processes of allo- and egocentric reference frame when switching between each other and contribute to our understanding of the interaction between two spatial reference frames. These studies suggest that we might systematically broaden our understanding of spatial reference frames by using the theory and methodology used in studies of task switching.

Asymmetrical Switch Costs

The task-switching paradigm has been extensively used in experimental psychology. In a typical task-switching experiment, identical or similar stimuli evoke distinct responses in different tasks. Participants are presented with a series of these stimuli and required to respond using the rules of different tasks, which switch frequently. Participants switch between the tasks flexibly; however, their performance is worse in switch trials (in which the current task differs from the previous one) than in repeat trials (in which the current task is the same as the previous one). Both higher response time (RT) and lower accuracy were found for the switch trials, which is called switch cost (e.g., Allport, Styles, & Hsieh, 1994; Jersild, 1927; Rogers & Monsell, 1995). Switch costs are generally attributed to two principal factors: One is the time required to reconfigure the changed task set (e.g., Rogers & Monsell, 1995; Rubinstein, Meyer, & Evans, 2001); the other is the interference from the previous task set (previous-trial interference), including the persisting suppression of the current task and the persisting activation of the previous task (e.g., Allport et al., 1994; Allport & Wylie, 1999).

The main evidence for the existence of previous-trial interference is the observed asymmetry in switch costs. Specifically, when participants switch to the stronger (dominant) task from the weaker (nondominant) one, switch costs are larger than when they switch to the weaker task from the stronger one (Allport et al., 1994; Allport & Wylie, 1999; Yeung & Monsell, 2003). For example, in Allport et al.’s (1994) study, participants were asked to switch between reading a colour word and naming the ink colour. The switch costs were higher for reading a word than for naming a colour. Allport et al. proposed an influential hypothesis of task-set inertia, which suggests that this asymmetry results from the asymmetry in the activation and inhibition of the two tasks, which are carried over from trial to trial (Allport et al., 1994; Allport & Wylie, 1999). In a task-switching paradigm, one needs to apply a strong inhibition to the dominant task set when performing the nondominant task (Allport et al., 1994), which carries over to the next trial. If a participant switches to the dominant task in the next trial, he or she needs to overcome this persisting inhibition (or negative priming), leading to a large switch cost. In contrast, performing the dominant task does not require a strong inhibition of the nondominant task set. Consequently, when switching to the nondominant task, little inhibition needs to be overcome, and there should be smaller switch cost. A number of studies showed that switch costs are usually higher for the more dominant task when participants switch between two tasks of unequal strength (for review on this issue, refer to Koch, Gade, Schuch, & Philipp, 2010). Such asymmetrical switch costs have been observed with different types of stimuli and response set (e.g., Allport & Wylie, 1999; Mueller, Swainson, & Jackson, 2009).

Study Objectives and Prediction

People usually switch from one reference frame to the other. Therefore, it would be worth understanding if and how two frames influence each other when people switch between them. In this study, we aimed to compare the relative strength of allo- and egocentric reference frames and to explore the interaction between them occurring in the process of switching between them. We developed an allocentric and egocentric switching task (AEST), which was modified from the classic AET (Galati et al., 2000; Neggers et al., 2005, 2006). In this study, participants were presented with stimuli in which a vertical target bar was placed on a horizontal background bar and were required to respond either to the midsagittal plane of their bodies (egocentric task) or to the centre line of the background bar (allocentric task). Different from the original AET, these two tasks switched frequently so that participants have to switch between two different spatial reference frames. Previous studies showed that the interaction between allo- and egocentric reference frames is unidirectional with the former affecting the latter but not vice versa (Neggers et al., 2005, 2006; Ruotolo, van der Ham, Iachini, & Postma, 2011), implying that the strength of allocentric representations is stronger than that of egocentric representations. In the current study, we wanted to explore whether this result would be observed in the process of switching. Therefore, the first hypothesis is that, if the strength of allocentric representation is stronger than the egocentric one, the allocentric representation would be less affected by the egocentric one in the process of switching between them. It is expected that the difference between the congruent condition and the incongruent condition (congruency effect) is smaller in the allocentric task than in the egocentric task.

Previous task-switching studies indicated that asymmetrical switch costs would be observed when participants switch between two tasks of unequal strength (Allport et al., 1994; Allport & Wylie, 1999; Yeung & Monsell, 2003). These findings inspired us to ask whether the switch costs are asymmetrical when the strength of allo- and egocentric tasks is unequal. To be specific, if the allocentric reference frame is stronger than the egocentric one, one may need to apply a strong inhibition to the allocentric processing when performing an egocentric task, and this inhibition carries over to the next trial. When the participant switches to allocentric task in the next trial, he or she needs to overcome this persisting inhibition, leading a larger switch cost. In contrast, a strong inhibition is not required when performing the allocentric task. As a result, when switching to the egocentric task, little inhibition needs to be overcome, and there should be smaller switch cost. Therefore, the second hypothesis is that, if the strength of two spatial reference frames are unequal, asymmetrical switch costs should be observed. It is expected that switch costs are lager in the allocentric task than in the egocentric task.

Method

Participants

Sixty-six students were recruited through advertisements in a student forum of Beijing Normal University and paid for their participation. All participants were right-handed and native Chinese speakers, with normal or corrected-to-normal vision. Four participants were excluded from the data analyses because of low accuracy rates (i.e., each of these four participants performed at chance or lower than chance in at least one condition). The final sample comprised 62 participants (31 women, age range: 17–30 years, M_age = 21.40 ± 2.33 years). All participants provided written, informed consent. The participants were free to withdraw at any time. The study design was approved by the institutional review board of the Faculty of Psychology, Beijing Normal University.

Materials

Eight visual stimuli were developed. Each stimulus consisted of a green or red vertical bar (width × height = 0.68 × 2.46° of visual angle; 24 bits RGB colour coding of the green bar: 0, 255, and 0; the red bar: 255, 0, and 0) placed in front of a grey horizontal background bar (width × height = 9.59 × 0.68° of visual angle; 24 bits RGB colour coding: 160, 160, and 160). The vertical bar could be located at –1.33 or 1.33° relative to centre of the horizontal background bar and –1.33 or 1.33° relative to midsagittal plane of the participant’s body, forming four possible combinations (see Figure 1). In the congruent condition, the vertical bar was on the same side with respect to the centre of the horizontal bar and the midsagittal plane of the participant’s body. In the incongruent condition, it was on different sides with respect to these two references frames.

Figure 1.

The eight stimuli used in an allocentric (red vertical bar) and egocentric (green vertical bar) switching task.

Apparatus

The stimuli were presented on a 22-inch colour monitor with a resolution of 1,680 × 1,050 pixels and a refresh rate of 60 Hz. The experimental room was darkened; hence, participants could hardly use the edges of the screen as a clue to make judgements. Participants were seated 57 cm in front of the screen, rendering the effective screen size as 45.12 × 29.12° visual angle. Their head movements were minimised by a chin rest, which was adjusted so that the eyes were at the same level as the vertical and horizontal centre of the screen.

Procedure

In each trial, the cue-stimulus interval was set at 0 ms and response-stimulus interval at 200 ms to induce sufficient switch costs (for a review, see Koch et al., 2010). To ensure that the participants used the midsagittal plane of their own body rather than other references in the egocentric tasks, no central fixation was presented. Each trial began with a stimulus displayed for 200 ms. After the stimulus disappeared, the participant had a maximum of 2 s to respond. If participants did not respond within 2 s, another blank screen appeared for 200 ms and then the next trial started.

The task was to decide whether the vertical bar was positioned on the left or on the right with respect to the different reference frames. For the green vertical bar, participants were instructed to judge its position with respect to their subjective body midline (egocentric task). For the red vertical bar, participants were instructed to judge its position with respect to the centre of the horizontal background bar (allocentric task). All participants responded using the same two keys on a keyboard—using “F” to signal left and “J” to signal right. Both speed and accuracy were emphasised.

There were two types of trials: switch and repeat trials. In switch trials, the reference frame used in the current trial differed from the previous trial (allocentric to egocentric task or vice versa). In repeat trials, the reference frame used in the current trial was the same as in the previous trial (allocentric to allocentric task or egocentric to egocentric task). There were four blocks in the experiment, with 66 trials in each block. All trial orders (i.e., Ego–Ego, Allo–Allo, Allo–Ego, and Allo–Ego) occurred equally often. In addition, half of the trials were congruent, and the other half were incongruent. All stimuli were presented in a pseudorandom order. Participants performed 34 practice trials and had to reach 90% accuracy to enter the formal experiment. There were 264 trials for each participant in the formal experiment.

Data Analysis

RTs and accuracy rates were analysed separately. The first two trials of each block were not analysed. In the RT analyses, the error and posterror trials (14.47%) and RTs beyond M ± 3 SD (1.37%) were excluded. A three-way repeated-measures analysis of variance was conducted with three within-participant factors: congruency (congruent vs. incongruent), spatial task (egocentric vs. allocentric judgement task), and trial type (switch vs. repeat).

Results

RT Analysis

Table 1 summarises the participants’ mean RTs and accuracies in each condition. The significant main effect of congruency, F(1, 61) = 45.62, p < .001, $η_{p}^{2}$ = .43, indicates that the RTs were lower (M = 643 ms) in the congruent trials than in the incongruent trials (M = 690 ms). The significant main effect of spatial task, F(1, 61) = 15.13, p < .001, $η_{p}^{2}$ = .20, indicated that the RTs in the egocentric task conditions (M = 685 ms) was higher than in the allocentric task conditions (M = 647 ms). The significant main effect of trial type, F(1, 61) = 371.93, p < .001, $η_{p}^{2}$ = .86, indicated that switch costs RTs of switch trials (M = 746 ms) were higher than that in repeat trials (M = 587 ms). In addition, the two-way interaction between congruency and spatial task was significant, F(1, 61) = 7.33, p < . 01, $η_{p}^{2}$ = .11, and the interaction between congruency and trial type was significant, F(1, 61) = 4.79, p = .03, $η_{p}^{2}$ =.07, but not for spatial task and trial type, F(1, 61) = 2.16, p = .15, $η_{p}^{2}$ = .03. The three-way interaction of congruency, trial type, and spatial task was also significant, F(1, 61) = 10.43, p < .01, $η_{p}^{2}$ = .15.

Table 1.

Mean Response Times (ms) and Accuracy for Repeat and Switch Trials in Allocentric and Egocentric Task.

Congruency	Task	Repeat		Switch		Cost^a
Congruency	Task	RT (ms)	Accuracy (%)	RT (ms)	Accuracy (%)	RT (ms)	Accuracy (%)
Congruent	Allocentric	549 (15.85)	97.68 (0.6)	714 (17.33)	95.77 (0.6)	165 (12.18)	1.9 (0.5)
Congruent	Egocentric	567 (18.73)	95.72 (0.6)	742 (18.75)	93.60 (0.9)	175 (12.96)	2.1 (0.6)
Incongruent	Allocentric	576 (16.83)	93.25 (0.7)	751 (15.32)	89.87 (0.9)	175 (14.64)	3.4 (0.9)
Incongruent	Egocentric	654 (17.65)	86.29 (1.5)	777 (17.17)	79.39 (1.7)	122 (11.38)	6.9 (1.8)

Note. RT = response time.

^aThe switch cost is the difference of RTs between repeat and switch trials.

To clarify the congruency effect, two subsequent separate analyses with congruency and spatial task for repeat and switch conditions (adjustment of Bonferroni) were conducted. For the repeat condition, both the main effect of congruency, F(1, 61) = 62.36, p < . 001, $η_{p}^{2}$ = .51, and the main effect of spatial task was significant, F(1, 61) = 13.08, p < . 001, $η_{p}^{2}$ = .18. The significant interaction, F(1, 61) = 17.15, p < . 001, $η_{p}^{2}$ = .22, indicated that the congruency effect was smaller for allocentric task (M = 26 ms) than for egocentric task (M = 87 ms; see Figure 2). For the switch condition, the significant main effect of congruency, F(1, 61) = 14.81, p < . 001, $η_{p}^{2}$ = .20, indicated that RTs were higher in the incongruent trials than in the congruent trials, and the significant main effect of spatial task, F(1, 61) = 6.58, p = . 01, $η_{p}^{2}$ = .10, indicated that RTs were higher for the egocentric task than for the allocentric task. The interaction between congruency and spatial task was not significant, F(1, 61) = 0.01, p = . 91, $η_{p}^{2}$ < .001. Those findings showed that there were significant congruency effects observed in both allo- and egocentric tasks and that the congruency effect in the egocentric task was larger than in the allocentric task for the repeat condition.

Figure 2.

For the repeat condition, the magnitude of the congruency costs (the difference between congruent and incongruent condition) in allo- and egocentric tasks. Error bars correspond to standard errors. Allo = allocentric task; Ego = egocentric task.

To explore the asymmetry of switch costs, two subsequent separate analyses for the incongruent and congruent conditions (adjustment of Bonferroni) were conducted. For the incongruent condition, the interaction between spatial task and trial type was significant, F(1, 61) = 9.96, p < .01, $η_{p}^{2}$ = .14, indicating that the switch cost was higher in the allocentric task (M = 175 ms) than in the egocentric task (M = 122 ms). A further simple effect analysis showed that switch costs appeared both in the allocentric task, F(1, 61) = 143.11, p < .001, $η_{p}^{2}$ = .70, and the egocentric task, F(1, 61) = 116.80, p < .001, $η_{p}^{2}$ = .62; see Figure 3). For the congruent condition, in contrast, the interaction between spatial task and trial type was not significant, F(1, 61) = .32, p = .57, $η_{p}^{2}$ = .01, indicating that the switch cost was symmetrical.

Figure 3.

Switch costs for allo- and egocentric task (the difference between switch and repeat condition) in congruent and incongruent trials. Error bars correspond to standard errors. Allo = switch to allocentric task; Ego = switch to egocentric task.

Accuracy Analysis

The accuracy results were similar to the results of RT analysis. The significant main effect of congruency, F(1, 61) = 93.08, p < .001, η2 p = .60, indicated that participants performed congruent trials (M = 95.68%) more accurately than incongruent trials (M = 87.20%). The significant main effect of spatial task, F(1, 61) = 45.63, p < .001, $η_{p}^{2}$ = .43, indicated that participants performed the allocentric task (M = 94.14%) more accurately than the egocentric task (M = 88.75%). The significant main effect of trial type, F(1, 61) = 43.33, p < .001, $η_{p}^{2}$ = .42, indicated switch costs: The accuracy was lower in switch trials (M = 89.66%) than in repeat trials (M = 93.24%). The two-way interaction between congruency and spatial task was significant, F(1, 61) =17.08, p < .001, $η_{p}^{2}$ = .22. The interaction between congruency and trial type was significant, F(1, 61) = 8.22, p < .01, $η_{p}^{2}$ = .12, but the interaction between spatial task and trial type was not, (F(1, 61) =2.76, p = . 10, $η_{p}^{2}$ = . 04. The three-way interaction of congruency, spatial task, and trial type was also not significant, F(1, 61) =2.17, p = . 15, $η_{p}^{2}$ = . 03.

To clarify the congruency effect, a two-way analysis of variance analysis with congruency and spatial task showed that both the main effect of congruency, F(1, 61) = 93.10, p < . 001, $η_{p}^{2}$ = .60, and the main effect of spatial task was significant, F(1, 61) = 45.63, p < . 001, $η_{p}^{2}$ = .43. The interaction between congruency and spatial task was significant, F(1, 61) = 17.08, p < . 001, $η_{p}^{2}$ = .22, indicating that the congruency effect is smaller in the allocentric task (M = 5.16%) than in the egocentric task (M = 11.82%). A further analysis showed that accuracy was lower in the incongruent condition than in the congruent condition, both in the egocentric task, F(1, 61) = 56.66, p < .001, $η_{p}^{2}$ = .48, and in the allocentric task, F(1, 61) = 70.18, p < .001, $η_{p}^{2}$ = .54. These results indicated again there were significant congruency effects observed in both allo- and egocentric tasks and that the congruency effect in the egocentric task was larger than in the allocentric task.

Discussion

This study investigated the relative strength of two spatial reference frames in switching processes and explored the switch costs and the asymmetry of switch costs in allo- and egocentric reference frames switching. First, higher RTs and lower accuracy were found in the incongruent condition than in the congruent condition for both allo- and egocentric tasks, but this effect was larger in the egocentric task than in the allocentric task. Second, there were salient switch costs when participants switch between allo- and egocentric reference frames. Both higher RTs and lower accuracy were observed in switch trials than in repeat trials. More important, the asymmetry of switch costs was observed in the incongruent condition. Specifically, the magnitude of switch costs was higher when switching from an egocentric task to an allocentric task than vice versa.

Previous studies (Neggers et al., 2005, 2006) found that egocentric presentations were affected by allocentric ones when participants performing allo- or egocentric task repeatedly, but not vice versa. To compare with previous studies, we investigated the congruency effect in the repeat condition. The congruency effect was larger in the egocentric task than in the allocentric task, indicating that the allocentric task was less susceptible to be interfered with by egocentric representations and thus superior in strength. These findings indicated that the allocentric reference frame was stronger than the egocentric one, which supported the first hypothesis. Moreover, the congruency effect may also provide new evidence pertaining to the two-visual-systems hypothesis (Goodale & Milner, 1992; Milner & Goodale, 2008; Ruotolo, van der Ham, Postma, Ruggiero, & Iachini, 2015). This hypothesis suggests that allocentric and egocentric reference frames have distinct functional roles within visuoperceptual and visuomotor tasks. Specifically, a visuomotor and visuoperceptual response favour egocentric and allocentric reference frames, respectively. Similar to previous studies (Neggers et al., 2005, 2006), the response modality in the current study might have triggered the participants to pay more attention to the perception-related than the motor-related components of the stimulus. As a result, an allocentric judgement was more advantageous than an egocentric one.

Interestingly, slightly different from previous findings (Neggers et al., 2005, 2006), the interaction was not completely unidirectional between the two reference frames in the present experiment. The patterns of RT and accuracy convergently showed that, not only was the egocentric task affected by the incongruent allocentric representation but also vice versa. A potential reason why previous studies (Neggers et al., 2005, 2006) did not show the bidirectional interaction might be the increased difficulty in our experiment. The task switching within a block required more cognitive resources than that in previous studies, in which participants did the same task consistently within a block. This might provide more chance for interference from egocentric presentations.

The switch costs in the allocentric task were larger than in the egocentric task, which verified the first hypothesis and supported the second hypothesis. This finding was consistent with previous findings (Neggers et al., 2005, 2006), which proposed that object position representations were stronger (more dominant) in the allocentric reference frame than in the egocentric reference frame; consequently, allocentric representations had a great influence on egocentric representations (Neggers et al., 2005). In the present study, according to the task-set inertia hypothesis, when one performed the less dominant of two tasks, he or she must inhibit the dominant task set that is a strong competitor. By contrast, when one performed the more dominant of two tasks, the inhibition of nondominant task set was less required. Hence, switching from egocentric (nondominant) to allocentric (dominant) task caused larger switch costs, because allocentric task was inhibited, which provoked by previous performing egocentric task. By contrast, when switching from an allocentric to egocentric task, switch costs are smaller due to less inhibition provoked by previous performing allocentric task.

Finally, it is worth noting that there are two limitations to the current study. In the present experiment, only one cue (colour) was used. Therefore, there might be some noise in the data because the cue-repetition effect might exaggerate the advantage of repeat trials (Logan & Bundesen, 2003; Mayr & Kliegl, 2003; Monsell & Mizon, 2006). Further research could use multiple cue-sets to eliminate this effect. Another limitation is that the colour-judgement assignment was not counterbalanced between participants, which might have caused an inequity between the two task difficulties and then influenced the asymmetry of switch costs. This limitation needs to be taken care of in future research.

In conclusion, the present study provides the first direct evidence that the strength of spatial reference frame plays an important role in spatial task switching, and the stronger allocentric reference frame leads a larger switch cost.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Fundamental Research Funds for the Central Universities (SKZZY2014055).

ORCID iD

Ming Lv

References

Acredolo

L. P.

(1978). Development of spatial orientation in infancy. Developmental Psychology, 14, 224–234.

Allport

Styles

E. A.

Hsieh

(1994). Shifting intentional set: Exploring the dynamic control of tasks. In Umilta

Moscovitch

(Eds.), Attention and performance XV: Conscious and nonconscious information processing (pp. 421–452). Cambridge, MA: MIT Press.

Allport

Wylie

(1999). Task-switching: Positive and negative priming of task-set. In Humphreys

G. W.

Duncan

Treisman

A. M.

(Eds.), Attention, space, and action: Studies in cognitive neuroscience (pp. 273–296). Oxford, England: Oxford University Press.

Bohbot

V. D.

Iaria

Petrides

(2004). Hippocampal function and spatial memory: Evidence from functional neuroimaging in healthy participants and performance of patients with medial temporal lobe resections. Neuropsychology, 18, 418–425.

Bridgeman

Peery

Anand

(1997). Interaction of cognitive and sensorimotor maps of visual space. Perception and Psychophysics, 59, 456–469.

Burgess

(2006). Spatial memory: How egocentric and allocentric combine. Trends in Cognitive Sciences, 10, 551–557.

Chen

Monaco

Byrne

Yan

Henriques

D. Y.

Crawford

J. D.

(2014). Allocentric versus egocentric representation of remembered reach targets in human cortex. Journal of Neuroscience, 34, 12515–12526.

Foley

R. T.

Whitwell

R. L.

Goodale

M. A.

(2015). The two-visual-systems hypothesis and the perspectival features of visual experience. Consciousness and Cognition, 35, 225–233.

Galati

Lobel

Vallar

Berthoz

Pizzamiglio

Le Bihan

(2000). The neural basis of egocentric and allocentric coding of space in humans: A functional magnetic resonance study. Experimental Brain Research, 133, 156–164.

10.

Galati

Pelle

Berthoz

Committeri

(2010). Multiple reference frames used by the human brain for spatial perception and memory. Experimental Brain Research, 206, 109–120.

11.

Goodale

M. A.

Milner

A. D.

(1992). Separate visual pathways for perception and action. Trends in Neurosciences, 15, 20–25.

12.

Jersild

A. T.

(1927). Mental set and shift. Archives of Psychology, 89.

13.

Klatzky

R. L.

(1998). Allocentric and egocentric spatial representations: Definitions, distinctions, and interconnections. In Freksa

Habel

Wender

K. F.

(Eds.), Spatial cognition: An interdisciplinary approach to representing and processing spatial knowledge (pp. 1–17). New York, NY: Springer-Verlag.

14.

Koch

Gade

Schuch

Philipp

A. M.

(2010). The role of inhibition in task switching: A review. Psychonomic Bulletin and Review, 17, 1–14. https://doi.org/10.3758/PBR.17.1.1

15.

Landau

Spelke

(1988). Geometric complexity and object search in infancy. Developmental Psychology, 24, 512–521.

16.

Logan

G. D.

Bundesen

(2003). Clever homunculus: Is there an endogenous act of control in the explicit task-cuing procedure? Journal of Experimental Psychology: Human Perception and Performance, 29, 575–599. https://doi.org/10.1037/0096-1523.29.3.575

17.

Mayr

Kliegl

(2003). Differential effects of cue changes and task changes on task-set selection costs. Journal of Experimental Psychology: Learning, Memory, and Cognition, 29, 362.

18.

Milner

A. D.

Goodale

M. A.

(2008). Two visual systems re-viewed. Neuropsychologia, 46, 774–785.

19.

Monsell

Mizon

G. A.

(2006). Can the task-cuing paradigm measure an endogenous task-set reconfiguration process? Journal of Experimental Psychology: Human Perception and Performance, 32, 493–516.

20.

Mou

McNamara

T. P.

Valiquette

C. M.

Rump

(2004). Allocentric and egocentric updating of spatial memories. Journal of Experimental Psychology: Learning, Memory, and Cognition, 30, 142–157. https://doi.org/10.1037/0272-7393.30.1.142

21.

Mueller

S. C.

Swainson

Jackson

G. M.

(2009). ERP indices of persisting and current inhibitory control: A study of saccadic task switching. NeuroImage, 45, 191–197.

22.

Nardini

Atkinson

Burgess

(2008). Children reorient using the left/right sense of coloured landmarks at 18-24 months. Cognition, 106, 519–527.

23.

Nardini

Burgess

Breckenridge

Atkinson

(2006). Differential developmental trajectories for egocentric, environmental and intrinsic frames of reference in spatial memory. Cognition, 101, 153–172.

24.

Neggers

S. F.

Scholvinck

M. L.

Van der Lubbe

R. H.

Postma

(2005). Quantifying the interactions between allo- and egocentric representations of space. Acta Psychologica, 118, 25–45.

25.

Neggers

S. F.

Van der Lubbe

R. H.

Ramsey

N. F.

Postma

(2006). Interactions between ego- and allocentric neuronal representations of space. Neuroimage, 31, 320–331.

26.

O’Keefe

Nadel

(1978). The hippocampus as a cognitive map. Oxford, England: Clarendon Press.

27.

Parslow

D. M.

Rose

Brooks

Fleminger

Gray

J. A.

Giampietro

, Brammer, M. J., Williams, S., Gasston, D., Andrew, C., Vythelingum, G. N., Loannou, G., Simmons, A., Morris

R. G.

(2004). Allocentric spatial memory activation of the hippocampal formation measured with fMRI. Neuropsychology, 18, 450–461.

28.

Roelofs

(1935). Optische localisation. Archive Für Augenheilkunde, 109, 395–415.

29.

Rogers

R. D.

Monsell

S. D.

(1995). Costs of a predictable switch between simple cognitive tasks. Journal of Experimental Psychology: General, 124, 207–231.

30.

Rubinstein

J. S.

Meyer

D. E.

Evans

J. E.

(2001). Executive control of cognitive processes in task switching. Journal of Experimental Psychology: Human Perception and Performance, 27, 763–797.

31.

Ruggiero

Iavarone

Iachini

(2018). Allocentric to egocentric spatial switching: Impairment in aMCI and Alzheimer’s disease patients? Current Alzheimer Research, 15, 1–8.

32.

Ruggiero

Ruotolo

Iachini

(2018). Congenital blindness limits allocentric to egocentric switching ability. Experimental Brain Research, 236: 813–820. https://doi.org/10.1007/s00221-018-5176-8

33.

Ruotolo

van der Ham

I. J. M.

Iachini

Postma

(2011). The relationship between allocentric and egocentric frames of reference and categorical and coordinate spatial information processing. The Quarterly Journal of Experimental Psychology, 64, 1138–1156.

34.

Ruotolo

van der Ham

I. J. M.

Postma

Ruggiero

Iachini

(2015). How coordinate and categorical spatial relations combine with egocentric and allocentric reference frames in a motor task: Effects of delay and stimuli characteristics. Behavioural Brain Research, 284, 167–178. https://doi.org/10.1016/j.bbr.2015.02.021

35.

Schenk

(2006). An allocentric rather than perceptual deficit in patient DF. Nature Neuroscience, 9, 1369–1370.

36.

Sterken

Postma

De Haan

E. H. F.

Dingemans

(1999). Egocentric and exocentric spatial judgements of visual displacement. The Quarterly Journal of Experimental Psychology, 52, 1047–1055.

37.

Vallar

Lobel

Galati

Berthoz

Pizzamiglio

Le Bihan

(1999). A fronto-parietal system for computing the egocentric spatial frame of reference in humans. Experimental Brain Research, 124, 281–286.

38.

Waller

Hodgson

(2006). Transient and enduring spatial representations under disorientation and self-rotation. Journal of Experimental Psychology: Learning, Memory, and Cognition, 32, 867–882.

39.

Yeung

Monsell

(2003). Switching between tasks of unequal familiarity: The role of stimulus-attribute and response-set selection. Journal of Experimental Psychology: Human Perception and Performance, 29, 455–469.