- Email: [email protected]
The influence of response grouping on free-choice decision making in a response selection task
Acta Psychologica 134 (2010) 175–181
Contents lists available at ScienceDirect
Acta Psychologica journal homepage: www.elsevier.com/locate/actpsy
The influence of response grouping on free-choice decision making in a response selection task Michael A. Khan a,*, Stuart Mourton a, Eric Buckolz b, Jos J. Adam c, Amy E. Hayes a a
School of Sport, Health and Exercise Sciences, Bangor University, UK School of Kinesiology, University of Western Ontario, Canada c Department of Movement Sciences, Maastricht University, The Netherlands b
a r t i c l e
i n f o
Article history: Received 22 September 2009 Received in revised form 25 January 2010 Accepted 26 January 2010 Available online 25 February 2010 PsycINFO classification: 2330 2346 Keywords: Response cuing Free choice Response selection Response grouping
a b s t r a c t Previous research has demonstrated an advantage for the preparation of fingers on one hand over the preparation of fingers on two hands, and for the preparation of homologous fingers over that of nonhomologous fingers. In the present study, we extended the precuing effects observed with finger responses to response selection under free-choice conditions. Participants were required to choose from a range of possible responses following the presentation of a precue that indicated which response to prepare (go-to precue) or prevent (no-go-to precue). In Experiment 1 the choice was between homologous and non-homologous finger responses on the hand opposite to the precue while in Experiment 2 the choice was between finger responses on the same or different hand to the precue. In the go-to precue condition, the frequency of homologous finger choices was more frequent than non-homologous finger responses. Similarly, participants chose finger responses on the same hand as the precue regardless of whether they were instructed to prepare or prevent the precued response. The hand effect bias was stronger than the finger effect bias. These findings are consistent with the Grouping Model (Adam, Hommel, & Umilta, 2003). Ó 2010 Elsevier B.V. All rights reserved.
1. Introduction In order to interact with a complex environment, people will allocate attention selectively to events that they believe to be most pertinent. It is well known that reaction times to predicted stimuli are faster than to those that are unexpected (Posner, 1980). For example, a tennis player would react quicker to a shot to their forehand if it was expected than if they predicted a shot to their backhand. The influence of advance information on reaction time has numerous practical applications for strategy and tactics in sport and combat situations. From a theoretical standpoint, it also provides a basis for understanding the processes underlying selective attention and movement preparation. Interestingly, while most research on selective attention has been directed towards understanding how excitatory and inhibitory processes influence latency measures, relatively little attention has been given to how these processes influence which response is actually chosen. In real world situations (e.g., sport, combat and driving), people are often placed in conditions of free choice where they have a
* Corresponding author. Address: School of Sport, Health and Exercise Sciences, Bangor University, George Building, Bangor, Gwynedd, Wales LL57 2PZ, UK. Tel.: +44 (01248) 388275; fax: +44 (01248) 371053. E-mail address: [email protected] (M.A. Khan). 0001-6918/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.actpsy.2010.01.008
certain degree of freedom to use an internal mode of selection to choose from a range of possible responses. In the present research, we consider how advance information and the mechanisms underlying selective attention influence selection processes when participants have the freedom to choose a response option from several alternatives (i.e., free choice). A common technique used to investigate mechanisms of selective attention and movement preparation is the precue paradigm. Precues give participants advance information about features of an upcoming stimulus and/or required response. This advance information may be valid or invalid depending on whether stimulus/response features match that of the precue. The typical finding is that reaction time for valid precue trials is quicker compared to when no advance information is provided whereas there is a reaction time cost on invalid precue trials. These reaction time benefits and costs have served as a basis for testing theories of selective attention (e.g., Bekkering & Pratt, 2004; Egly, Driver, & Rafal, 1994) as well as processes underlying movement selection and preparation (e.g., Larish & Frekany, 1985; Rosenbaum, 1980). In most research on selective attention, either a many:1 or a 1:1 stimulus/response mapping has been employed. In the case of many:1 mappings, the response is specified from the outset and hence response selection processes are not involved. For 1:1 stimulus/response mappings, the response to be produced is fully
176
M.A. Khan et al. / Acta Psychologica 134 (2010) 175–181
specified by the characteristics of the stimulus. Therefore, regardless of whether one must react to a situation that is expected or unexpected, the selection of responses is externally determined since there is only one appropriate response on any particular trial. On the other hand, paradigms involving a 1:many stimulus/response mapping allow participants to make internal decisions when choosing freely from a subset of possible responses. It has been suggested that mechanisms of selection that underlie decision making in internal versus external choice situations are fundamentally different (e.g., Baylis, Tipper, & Houghton, 1997; Buckolz, Goldfarb, & Khan, 2004; Keller et al., 2006). However, it is interesting to note that most research on attention and speeded decision making has employed forced choice situations in which the selection of responses is externally determined. Although the study of how advance information influences speeded free choice is relatively limited, it has been demonstrated that the presentation of a subliminal cue influences response selection (Klapp & Haas, 2005; Schlaghecken & Eimer, 2004). In these studies, participants were first presented with a masked arrow that pointed left or right. This was followed by a stimulus arrow that pointed either to the left or right (forced choice trials) or a double headed arrow (free choice trials). It was demonstrated that on free choice trials, responses that corresponded to the precue were produced more often when the interval between the precue and the stimulus was short. At longer precue-stimulus intervals, there was a bias towards selecting responses that were opposite to the precue. Thus, it appears that the presentation of subliminal precues leads to activation and subsequent inhibition of motor responses resulting in systematic response selection biases. Posner, Rafal, Choate, & Vaughan (1985) have provided similar evidence that inhibition influences free choice by demonstrating that the selection of eye movements is biased away from locations that have been associated with earlier inhibition. The goal in the present research was to examine whether the instruction to prepare or prevent a precued response would influence which response option is chosen from a subsequently specified range of alternatives. The rationale behind our investigation was that when participants are instructed to prepare a particular response, the activation of this response will bias selection towards alternatives that share similar features (e.g., the hand or homologous finger). Conversely, when participants are instructed to prevent a particular course of action, the inhibition of a response option might bias response selection away from alternatives that share similar features. An underlying assumption is that response selection biases will arise from the grouping of response features between sets of alternatives. Strong evidence in support of this assumption stems from studies that have used the finger-cuing paradigm to investigate the efficiency of finger grouping processes. The finger-cuing paradigm was developed by Miller (1982), who adapted Rosenbaum’s movement precuing technique (Rosenbaum, 1980, 1983). In the finger-cuing task, a visual cue signal temporally precedes the target signal. The cue specifies a subset of two of four possible (keypress) responses (implemented by the index and middle fingers of both hands), thus prompting a process of subgroup making. In the hand-cued condition, the cue specifies two fingers on the same hand (e.g., the left-index finger and the left-middle finger). In the finger-cued condition, the cue specifies the homologous fingers on different hands (e.g., the two index fingers). In the neither-cued condition, the cue specifies non-homologous fingers on different hands (e.g., the left-middle and rightindex fingers). Also, a neutral (control) condition is included, which provides no advance information, and thus precludes the grouping and preparation of any combination of two finger responses. Cue effectiveness is inferred from a significant RT advantage for the informative cue conditions (i.e., hand-cued, finger-cued, and neither-cued) over the control, uninformative (neutral) cue condition.
The consistent finding from the finger-cuing paradigm is a pattern of differential cuing benefits that is apparent with short preparation intervals (i.e., intervals shorter than about 2–3 s). RTs are shortest for the hand-cued condition and longest for the neithercued condition, with RTs for the finger-cued condition being intermediate (for reviews see Adam et al., 2003, 2005;, Reeve & Proctor, 1990). A recent account of this pattern of differential cuing benefits is the Grouping Model (Adam et al., 2003, 2005), which is an extension of the salient-features coding principle advanced by Proctor & Reeve (1988) & Reeve & Proctor (1990). The key idea of the Grouping Model is that the individual elements of multi-element visual displays and multi-element response arrays are not processed independently but are pre-attentively organized or ‘‘grouped” according to low-level grouping factors that depend on stimulus driven factors (e.g., Gestalt principles) and on response-related factors (e.g., inter-response linkages). Preattentive processing is done quickly, effortlessly and in parallel without the need of focused attention (cf. Treisman, 1986). In other words, the basic assumption is that, each stimulus set and each response set has a default organization established automatically by the bottom-up computation of perceptual and motoric units or subgroups; this process is fast and effortless. With additional, top-down processing, however, alternative organizations can be attained; this process is slow and effortful. Thus, the pattern of cuing effects that emerge in the finger-cuing task critically depends on the nature of these default groupings and on the time available to reorganize these representations, if necessary. According to the Grouping Model, the processing advantage of hand-cues simply reflects the natural, strong grouping of two fingers on the same hand. The co-activation of directly adjacent and overlapping cortical finger representations in areas of the motor cortex corresponding to the same-hand finger set could be at the basis of this grouping (e.g., Dechent & Frahm, 2003). The bilateral finger- and neither-cues, on the other hand, are more difficult to process because they require slow, effortful, top-down modulation to breakup the anatomically based left–right motor organization and to create a new motor organization represented in two hemispheres. Moreover, the advantage of finger-cues over neither-cues can be attributed to the fact that finger-cues require the grouping of homologous fingers whereas neither-cues require the grouping of two different, non-homologous fingers. Because homologous fingers are neurally and functionally linked (e.g., Stinear, Walker, & Byblow, 2001; Ugawa, Hanajima, & Kanazawa, 1993), they are easier to group or co-activate than non-homologous fingers. Hence, the existence of facilitatory connections between homologous motor areas in the brain implies that the preparation and activation of one finger may spill over to its homologous counterpart in the opposite hemisphere. In the present experiments, we extended the precuing effects observed with finger responses to response selection under freechoice conditions. A novel paradigm was employed in which participants were required to choose from a range of possible responses following the presentation of advance information. On each trial, a precue was presented that indicated to the participant which response to prepare (i.e., go-to precue). This was followed by the imperative stimulus that specified a subset of possible responses rather than a specific response (1: many stimulus–response mapping). In Experiment 1, the imperative stimulus specified the finger responses on either the left or right hand. When the precued response was on the same hand specified by the imperative stimulus, participants were required to produce the precued response. However, if the imperative stimulus specified the other hand, participants were placed in a free-choice situation. Hence, they could choose between a homologous and nonhomologous finger relative to the finger specified by the precue. Based on the above reviewed evidence that supports homologous
M.A. Khan et al. / Acta Psychologica 134 (2010) 175–181
finger grouping over non-homologous finger grouping, we expected participants in a free-choice situation to opt more often for a homologous than for a non-homologous finger response. We also employed trials in which participants were instructed to prevent the precued response (no-go-to precue). In these cases, it was expected that the inhibition of the precued response would bias choices away from a homologous finger response in the free-choice situation. 2. Experiment 1 2.1. Method 2.1.1. Participants Twenty-four self declared, right hand dominant, university students served as participants in the study (12 males, 12 females, ages 18–35 years). All participants in this Experiment and Experiment 2 were naive to the hypothesis being tested and inexperienced at the experimental task. They gave their informed consent prior to participation in the study. The experimental protocols were approved by the Ethics Committee of the School of Sport, Health and Exercise Sciences, Bangor University for research involving human participants in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki. 2.1.2. Apparatus Participants sat in a chair in front of a computer monitor that was placed on a table top. The monitor was approximately 45 cm from the eyes of the participants. The display on the monitor consisted of four squares (4 4 cm) arranged horizontally with a separation of 4 cm (side to side) on a black background. The squares had a white outline and were not filled. The middle and index fingers of the left hand were placed on the ‘‘F’ and ‘‘G” keys, respectively, while the index and middle fingers of the right hand were placed on the ‘‘J” and ‘‘K” keys, respectively. The four squares on the monitor were mapped in a spatially compatible manner with the four finger presses such that the leftmost square corresponded to the left-middle finger keypress, and so on. 2.1.3. Procedure At the beginning of each trial, the four target squares appeared on the screen. After a delay of 2000 ms a precue was presented by coloring green the outlines of one of the squares for a duration of 1000 ms. Then, following an interval of either 100 or 750 ms, the imperative stimulus was presented between the second and third squares. Thus, there were two stimulus-onset asynchronies (SOAs) between precue and imperative stimulus onset, namely 1100 and 1750 ms. The imperative stimulus consisted of an arrow pointing to the right or left (see Fig. 1). The imperative stimulus arrow defined a subset of two responses. An arrow pointing to the right specified the two right hand finger responses while an arrow pointing to the left corresponded to the two left hand responses.
Precue
Stimulus Forced-choice
177
The imperative stimulus remained on screen until the participant made a response. In one block of trials (i.e., go-to precue), participants were instructed to press the response corresponding to the precue if the stimulus arrow pointed to the same side as the precue. This was a forced choice situation in that only one alternative was appropriate. For example, if the right-index finger was precued and the imperative stimulus arrow pointed to the right, participants had to depress the right-index finger (i.e., J key). However, if the arrow pointed to the opposite side to that of the precue, the participant now had a free-choice situation in that they could choose either of the two responses on the left hand. In another block of trials (no-go-to precue), participants were told to prevent making the response that corresponded to the precue. Therefore, if the stimulus arrow pointed to the same side as the precue, participants were required to produce the response on that hand that did not correspond to the precue. For example, if the right-index finger was precued and the arrow pointed to the right, participants were required to respond with their right middle finger. If the arrow pointed to the left, participants again had a free-choice situation and could choose between the index and middle fingers of the left hand. Each participant performed two blocks of trials, one in which the precues were designated as go-to precues and one in which precues were designated as no-go-to precues. For each block of trials, the eight combinations of the four precue positions and two imperative stimulus directions were randomized in a pseudo random fashion such that each combination was administered before any was repeated. Each combination was administered twenty times giving a total of 160 trials per block. Since each combination of precue position and stimulus direction occurred equally often, there was a probability of 50% that the imperative stimulus would point to the same or opposite side to that of the precue. The order of the go-to and no-go-to precue trial blocks was counterbalanced between participants. 2.1.4. Analytic methods Of primary interest in the present experiment were those trials in which the stimulus arrow pointed to the opposite direction to that of the precue (i.e., free choice trials). Specifically, we were interested in whether participants would choose the finger response that was homologous or non-homologous to the precue. For the free choice data, since the homologous finger condition and the non-homologous finger condition are interdependent, they cannot be analysed as independent levels of a Choice factor. Hence, to determine whether one finger response was produced more often than the other, the difference in response frequencies was calculated between the homologous finger and non-homologous finger values for both the go-to and no-go-to conditions. One-sampled t-tests were then used to test whether this difference was significantly greater than zero. A preliminary analysis revealed that these difference scores were not affected by SOA for either Precue Condition, and therefore the SOA factor will not be discussed further. Forced choice and free choice RTs were analysed using a 2 Precue Condition (go-to, no-go-to) 3 Choice (forced choice, homologous finger, non-homologous finger) fully repeated measures ANOVA. 2.2. Results and discussion
Free-choice LM
LI
RI
RM
Fig. 1. Sequence of events on each trial from precue to imperative stimulus presentation on forced and free choice trials in Experiment 1 (LM – left middle; Li – left index; RI – right index; RM – right middle). All squares were outlined squares, but are depicted as filled in for clarity.
2.2.1. Errors The overall error rate in the go-to trial blocks was 9.6% (see Table 1). On forced choice trials, the most frequent errors were those in which participants produced the incorrect finger response on the same hand as the precue. The most frequent errors on the free choice trials were those in which the precued response was
178
M.A. Khan et al. / Acta Psychologica 134 (2010) 175–181
Table 1 Percentage of errors for the go-to and no-go-to precues under the free and force choice conditions in Experiment 1. Errors are coded in relation to the precued finger (SH = same hand; DH = different hand; SF = same finger; DF = different finger). Go-to precue
No-go-to precue
Forced choice
Forced choice
SH/DF
DH/SF
DH/DF
5.6
0.6
0.1
Free choice
Precue
DH/SF
DH/DF
4.5
0.6
1.1
Free choice
Precue
SH/DF
Precue
SH/DF
2.9
0.4
0.4
1.4
produced. The overall error rate in the no-go-to trial blocks was 8.0%. The highest error rates were on forced choice trials in which the precued response was produced. Trials in which errors were committed were excluded from the following response frequency and reaction time analyses. 2.2.2. Response frequencies The proportion of free choice trials for which participants produced homologous and non-homologous finger responses to that of the precue are shown in Fig. 2. One-sampled t-tests revealed that for the go-to condition, the difference was greater than zero (t(23) = 3.50; p = 0.002), indicating that homologous finger responses were significantly more frequent than the non-homologous finger responses. Previous research has revealed that RTs were quicker when homologous finger pairs were precued compared to when nonhomologous fingers were precued (Adam et al., 2003, 2005). In those studies, two responses were specified by the precue while the imperative stimulus specified one of the two precued responses. The results of the current study in which participants were given a free choice following the presentation of a single response precue are consistent with these findings. The finding that participants chose the finger response that was homologous to the precue more often than the non-homologous finger indicates that the preparation of a single response does bias decision making towards alternatives that share similar features (i.e., homologous finger). For the no-go-to condition, the difference between the homologous and non-homologous finger responses was not significantly different from zero (t(23) = 0.26; p = 0.8). We had expected that the instruction to prevent a particular response would bias deci-
homologous non-homologous
0.8
Response Frequency
2.2.3. Reaction times The analysis of RT revealed a significant Precue Condition Choice interaction, F(2, 46) = 66.2, p < 0.01. Breakdown of this interaction using Tukey HSD post hoc tests (p < 0.05) revealed that in the go-to condition, RTs were shorter for the forced choice compared to free choice trials, a finding that reflects the instruction to prepare the precued response (see Table 2). Also, on free choice trials, RTs were shorter for homologous finger than for non-homologous finger responses. Hence, not only did participants choose homologous finger responses more often, but RTs were shorter for these responses compared to non-homologous fingers. In the no-go-to condition, RTs were longer in the forced choice condition compared to when a non-homologous response was chosen on free choice trials. This may reflect a strategy adopted by participants to avoid using the hand that contained the no-go-to precue. Similar to the response frequency results in the no-go-to condition, there was no difference in RT between the homologous and non-homologous responses on free choice trials. 3. Experiment 2 In Experiment 1, the imperative stimulus specified either the two left or right hand responses. Hence, on free choice trials, the choice was between homologous and non-homologous finger responses relative to the precue. In Experiment 2, the imperative stimulus specified the two index fingers or the two middle fingers. In free choice trials, the choice was between (non-homologous) finger responses located on the same or different hand relative to the finger specified by the cue. Based on evidence that within-hand grouping is much stronger than between-hands grouping, we expected participants in the free-choice situation to show a robust selection preference for the same-hand finger response. Furthermore, this selection bias for a same-hand over a different-hand response alternative should be stronger than the selection bias demonstrated in Experiment 1 for a homologous over a nonhomologous response alternative. This would reflect the dominance of hand grouping over (homologous) finger grouping. 3.1. Method 3.1.1. Participants Twenty-four self declared, right hand dominant, university students served as participants in the study (12 males, 12 females, ages 18–35 years).
1 0.9
sion making away from responses that shared similar features. However, this was not the case since the frequency of non-homologous finger responses was not significantly different from homologous responses.
3.1.2. Apparatus Same as in Experiment 1.
0.7 0.6
3.1.3. Procedure The procedures were similar to Experiment 1 except that the imperative stimulus now consisted of arrows pointing inwards or
0.5 0.4 0.3 0.2
Table 2 Mean (standard deviation) reaction times (ms) under go-to and no-go-to precue conditions in Experiment 1 for forced choice and free-choice conditions.
0.1 0 go-to
no-go-to
Precue Condition Fig. 2. Response frequencies for homologous and non-homologous finger options for the go-to and no-go-to precues in the free-choice condition of Experiment 1.
Response
Go-to precue
No-go-to precue
Forced choice Free choice: homologous finger Free choice: non-homologous finger
424 (67) 477 (64) 490 (69)
482 (69) 475 (50) 467 (53)
179
M.A. Khan et al. / Acta Psychologica 134 (2010) 175–181
3.1.4. Analytic methods The statistical analyses were analogous to those of Experiment 1. As in Experiment 1, initial analyses indicated no significant effects of SOA on response frequencies, and hence this factor will not be considered further.
3.2. Results and discussion 3.2.1. Errors The overall error rate in the go-to trial blocks was 8.8% (see Table 3). On forced choice trials, the highest errors rates occurred when participants produced the homologous finger response on the opposite hand to the precue. On free choice trials, the most frequent errors were those in which the precued finger response was produced. The overall error rate in the no-go-to trial blocks was 8.9%. On forced choice trials, the highest errors occurred when participants produced the precued response. On free choice trials, the most frequent errors were those in which the homologous finger response on the opposite hand to the precue was produced.
Precue
Stimulus
><
Forced-choice
<>
Free-choice
Table 3 Percentage of errors for the go-to and no-go-to precues under the free and force choice conditions in Experiment 2. Errors are coded in relation to the precued finger (SF = same finger; DF = different finger; SH = same hand; DH = different hand). Go-to precue
No-go-to precue
Forced choice
Forced choice
SF/DH
DF/SH
DF/DH
Precue
DFSH
DF/DH
2.8
0.8
0.1
2.7
0.4
1.9
Free choice
Free choice
Precue
SF/DH
Precue
SF/DH
4.9
0.2
0.1
3.0
Trials in which errors were committed were excluded from the following response frequency and reaction time analyses. 3.2.2. Response frequencies Of primary interest in the present experiment was whether participants would choose the finger response on the same or other hand to that of the precue on free choice trials. The proportion of free choice trials for which the participants produced a finger response on the same and different hand to that of the precue are shown in Fig. 4. Similar to Experiment 1, a difference score was calculated between the same and different hand frequencies for both the go-to and no-go-to conditions. One-sampled t-tests revealed that the difference was greater than zero for both the go-to condition (t(23) = 14.30; p < 0.001) and the no-go-to condition (t(23) = 10.41; p < 0.001). As shown in Fig. 4, responses on the same hand to that of the precue were chosen more often regardless of whether participants were instructed to prepare or prevent the precued response. A paired-samples t-test revealed that the tendency to select the same-hand finger response was significantly greater in the go-to condition than in the no-go-to condition (t(23) = 3.57; p = 0.002). This relatively large bias towards selecting a response from the same hand as the precue reflects the strong nature of within-hand grouping. Previous research has demonstrated that RTs were shortest when two responses on the same hand were precued compared to when two homologous or nonhomologous fingers on different hands were precued. The present results imply that this within-hand advantage also emerges in the form of response biases under free-choice conditions. 3.2.3. Reaction times As mentioned above, there was a strong tendency to choose responses on the same hand as the precue. Data from one participant
1
same hand different hand
0.9
Response Frequency
outwards (see Fig. 3). Arrows pointing inwards specified the index fingers of each hand while arrows pointing outwards specified the middle fingers. Therefore, in contrast to Experiment 1 where the imperative stimulus specified finger responses on the same hand, the imperative stimulus now specified homologous fingers on separate hands. Hence, in free-choice situations, the choice was in effect between hands instead of between fingers. Participants performed one block of trials under the go-to precue condition and another block under the no-go-to precue condition. As in Experiment 1, participants were instructed to prepare the precued response in the go-to condition. For example, if the right-index finger was precued and the stimulus arrows pointed inwards, participants were required to depress the right-index finger. However, if the arrows pointed outwards, participants could choose either the right or left-middle finger. In the no-go-to condition, participants were instructed to prevent making the response indicated by the precue. Hence, if the arrows pointed to a subset of finger responses that included the precued finger response, participants were required to produce the response opposite to the position of the precue. For instance, if the right-index finger was precued and the arrows pointed inwards, participants were required to respond with the left-index finger. However, if the arrows pointed outwards, participants were again free to choose between the middle finger of the left hand and the middle finger of the right hand. For each block of trials, the eight combinations of the four precue positions and two stimulus arrows were randomized in a pseudo random fashion such that each combination was administered before any was repeated. Each combination was administered twenty times giving a total of 160 trials per block. The order of the go-to and no-go-to precue trial blocks was counterbalanced between participants.
0.8 0.7 0.6 0.5 0.4 0.3 0.2
LM
LI
RI
RM
0.1 0
Fig. 3. Sequence of events on each trial from precue to stimulus presentation on forced and free choice trials in Experiment 2 (LM – left middle; Li – left index; RI – right index; RM – right middle). Stimuli were outlined squares, but are depicted as filled in for clarity.
go-to
no-go-to
Precue Condition Fig. 4. Response frequencies for same hand and different hand options for the go-to and no-go-to precues in the free-choice condition of Experiment 2.
180
M.A. Khan et al. / Acta Psychologica 134 (2010) 175–181
were excluded from the RT analysis because there were no different-hand responses in at least one of the experimental conditions. The analysis of RTs from the remaining 23 participants was performed using a 2 Precue Condition (go-to, no-go-to) 3 Choice (forced choice, same hand, different hand) ANOVA. This analysis revealed a significant Precue Condition Choice interaction, F(2, 44) = 39.6, p < 0.001. Breakdown of this interaction using Tukey HSD post hoc tests (p < .05) revealed that RTs in the go-to precue condition were faster for the forced compared to free choice trials (see Table 4). Also, when participants chose a response that was on the same hand as the precue, RTs were faster than when they chose a response on the other hand. This again demonstrates the strong influence of within-hand grouping on response latencies. There were no differences in RTs in the no-go-to condition. 3.2.4. Between experiment comparison In order to compare the magnitude of free choice biases between Experiment 1 and Experiment 2, the difference in frequencies between same and different finger responses in Experiment 1 and between same and different hand responses in Experiment 2 were submitted to a 2 Experiment 2 Precue (go-to, no-go-to) ANOVA with repeated measures on the second factor. A main effect of Experiment, F(1, 46) = 136, p < 0.001, indicated that the difference between same and different-hand response frequencies in Experiment 2 was greater than the difference between homologous and non-homologous finger frequencies in Experiment 1. Also, a main effect of Precue, F(1, 46) = 21.9, p < 0.001, indicated that response biases were greater in the go-to compared to nogo-to condition. These results further demonstrate stronger within compared to between hand grouping of finger responses. 4. General discussion Past research has shown that when two of four finger responses are precued, RTs are fastest when the two precued responses are on the same hand, slowest when the precue specifies non-homologous fingers from each hand, and intermediate RTs when homologous fingers are precued (Adam & Pratt, 2004). In the present study, we extended these precuing effects to free-choice situations. Our primary goal was to investigate if a precued response would bias the subsequent response selection processes. The results of Experiment 1 revealed that for go-to precues, participants selected the homologous finger responses in free-choice conditions more often than non-homologous responses. No-go-to precues did not have an influence on response selection. In Experiment 2, participants were given a choice between homologous finger responses on separate hands. There was a strong bias towards choosing responses on the same hand as the precue. Although this was the case for both go-to and no-go-to precues, the bias towards within-hand choices was stronger in the go-to condition. In order to account for the effects of advance information on RTs of finger responses, Adam et al. (2003, 2005) have proposed a Grouping Model that takes into account coding processes at both stimulus and response levels. According to the Grouping Model, arrays of stimuli and responses are separately organised pre-attentively into subgroups. In cases in which the precue defines a
Table 4 Mean (standard deviation) reaction times (ms) under go-to and no-go-to precue conditions in Experiment 2 for forced choice and free-choice conditions. Response
Go-to precue
No-go-to precue
Forced choice Free choice: same hand Free choice: different hand
475 (38) 547 (54) 574 (81)
536 (52) 542 (48) 548 (56)
subgroup of stimuli that maps directly to a subgroup of responses, a fast automatic selection of responses occurs. When the precue does not correspond to a subgroup of stimuli or there is not a direct relation between stimulus and response subgroups, effortful processing is needed to select and activate responses. The Grouping Model has been based predominantly on experiments in which the precue specifies a subset of responses (e.g., two of four finger keypresses) from which one would be specified by the imperative stimulus. Hence, the efficiency of subgroup formation is contingent upon which responses are specified by the precue. The paradigm in the present study differs from past work in that the precue specified only one response (not two). Participants were then required to choose from a subset of responses that was specified by the imperative stimulus. Hence, in contrast to the classic finger-precuing paradigm, the precue itself did not specify a subgroup of responses. Of interest was whether the preparation of a single response would lead to subgroup formation by co-activation of other responses that shared similar features with the precued response. If this were the case, response selection would be biased towards responses that were within the same subgroup as the precued response. More specifically, based on the assumption that fingers on the same hand form a strong, well-defined subgroup, the activation of one finger response would result in simultaneous activation of other fingers on the same hand. This increased activation would bias response selection in free-choice situations towards the more highly activated responses on the same hand as the precue compared to responses on the other hand. The findings of Experiment 2 substantiated this assumption by showing that participants strongly preferred a response option on the same hand as the precue compared to a response option on the other hand. Similarly, the activation of a single finger response may not only be limited to responses on the same hand but also to fingers on the other hand that share a common feature. This was demonstrated in Experiment 1 where response selection was biased towards homologous finger pairings on the opposite hand to that of the precue. Since both experiments revealed that the instruction to prepare a single response influenced free choice response selection, it appears that both within-hand and homologous finger pairings belong to well-defined (motor) subgroups that are neurally linked (e.g., Dechent & Frahm, 2003; Stinear et al., 2001). It is possible that participants may have adopted a strategic approach to subgroup formation upon presentation of the precue in order to reduce the number of alternatives. For example, in Experiment 1, when the right-index finger was precued, participants may have prepared the right-index finger (in the event of a forced choice stimulus) along with the left-index finger (in the event of a free choice stimulus). Similarly, in Experiment 2, when the rightindex finger was precued, participants may have simultaneously prepared the right middle finger. However, this does not appear to be the case since reaction times for the precued responses were clearly shorter than other responses. Hence, participants were preparing the precued response to a level beyond all other alternatives. In Experiment 1, participants were required to choose between homologous or non-homologous finger responses on the opposite hand to the precue. In Experiment 2, the choice was between finger responses on the same or opposite hand to the precue. Comparison between experiments indicated that fingers on the same hand constituted a stronger subgroup than homologous finger pairings on different hands. When participants were required to choose between finger responses on different hands, the tendency to choose a response on the same hand as the precue was stronger than choosing homologous finger responses on the hand that did not contain the precued response. These findings then are compatible with the Grouping Model since they conform to the pattern of
M.A. Khan et al. / Acta Psychologica 134 (2010) 175–181
differential precuing benefits where hand-cues are most effective, neither-cues (non-homologous fingers) least, and finger-cues (homologous fingers) intermediate, indicating that within-hemisphere co-activation of finger responses is stronger than between-hemisphere co-activation of finger responses. Although go-to precues had an influence on response frequencies and reaction times, the influence of no-go-to precues on response selection was relatively weak. Given the evidence for inhibitory connections within and between the two hemispheres (e.g., Ferbert et al., 1992), we had expected that the instruction to prevent a particular response would lead to the inhibition of other responses that shared similar features (i.e., responses within the same subgroup). However, the results of Experiment 1 did not reveal a bias away from homologous finger responses. In Experiment 2, on the other hand, there was a significant predisposition to choose responses on the same hand as the no-go-to precue but this bias was not as strong when compared to go-to precues. Hence, these findings do not unequivocally support our hypothesis that the prevention of a particular response biases selection away from other responses in the same subgroup. This would be in contrast to other studies that have shown that responses are biased away from the direction of a subliminal precue (i.e., masked primes) at long precue-stimulus ISIs (e.g., Schlaghecken & Eimer, 2004). It was proposed that the initial activation of the cued response was followed by inhibition that consequently led to a reduced frequency and increased RT for the cued response. However, one important difference in the design of our study was that the no-go-to response could not be used whereas it remained a viable option in past work. Hence, it may be that while the inhibitory effects associated with a cued response bias response selection away from that response, they are not of sufficient strength to transfer to other responses. It is also possible that the instruction to prevent a particular response was not a true test of how its inhibition influences other responses. This is because participants may have prepared alternative responses upon presentation of a no-go-to precue; such a recoding strategy would be akin to ‘‘indirect cuing” making it similar to the go-to condition. Hence, the results of the no-go-to trials should be interpreted with caution at this point. In summary, the present results offer strong support for the Grouping Model proposed by Adam et al. (2003, 2005). Past work has clearly demonstrated the benefit of subgroup formation on response preparation through the shortening of response latencies. In the present study, we extended this to free-choice situations. When participants were instructed to prepare a response, response selection under free-choice conditions was biased towards responses that shared similar features with the precued response. This finding is a powerful demonstration of subgroup formation (or response co-activation) even in cases in which only one response is specified by the precue.
181
References Adam, J. J., Hommel, B., & Umilta, C. (2003). Preparing for perception and action (I): The role of grouping in the response-cuing task. Cognitive Psychology, 46, 302–358. Adam, J. J., Hommel, B., & Umilta, C. (2005). Preparing for perception and action (II): Automatic and effortful processes in response cueing. Visual Cognition, 12, 1444–1473. Adam, J. J., & Pratt, J. (2004). Dissociating visual attention and effector selection in spatial precuing tasks. Journal of Experimental Psychology: Human Perception and Performance, 30, 1092–1106. Baylis, G. C., Tipper, S. P., & Houghton, G. (1997). Externally cued and internally generated selection: Differences in distractor analysis and inhibition. Journal of Experimental Psychology: Human Perception & Performance, 23, 1617–1630. Bekkering, H., & Pratt, J. (2004). Object-based processes in the planning of goaldirected hand movements. The Quarterly Journal of Experimental Psychology, 8, 1345–1368. Buckolz, E., Goldfarb, A., & Khan, M. A. (2004). The use of distractor assigned response slows later responding in a location negative priming task. Perception and Psychophysics, 66, 837–845. Dechent, P., & Frahm, J. (2003). Functional somatotopy of finger representations in human primary motor cortex. Human Brain Mapping, 18, 272–283. Egly, R., Driver, J., & Rafal, R. D. (1994). Shifting visual attention between objects and locations: Evidence from normal and parietal lesion subjects. Journal of Experimental Psychology: General, 123, 161–177. Ferbert, A., Priori, A., Rothwell, J. C., Day, B. L., Colebatch, J. G., & Marsden, C. D. (1992). Interhemispheric inhibition of the human motor cortex. Journal of Physiology, 453, 525–546. Keller, P. E., Wascher, E., Prinz, W., Waszak, F., Koch, I., & Rosenbaum, D. A. (2006). Differences between intention-based and stimulus-based actions. Journal of Psychophysiology, 20, 9–20. Klapp, S. T., & Haas, B. W. (2005). Nonconscious influence of masked stimuli on response selection is limited to concrete stimulus–response associations. Journal of Experimental Psychology: Human Perception and Performance, 31, 193–209. Larish, D. D., & Frekany, G. A. (1985). Planning and preparing expected and unexpected movements: Reexamining the relationships of arm, direction, and extent of movement. Journal of Motor Behavior, 17, 168–189. Miller, J. (1982). Discrete versus continuous models of human information processing: In search of partial output. Journal of Experimental Psychology: Human Perception and Performance, 8, 273–296. Posner, M. I. (1980). Orienting of attention. The VIIth Sir Frederic Bartlet Lecture. Quarterly Journal of Experimental Psychology, 32, 3–25. Posner, M. I., Rafal, R. D., Choate, L. S., & Vaughan, J. (1985). Inhibition of return: Neural basis and function. Cognitive Neuropsychology, 2, 211–228. Proctor, R. W., & Reeve, T. G. (1988). The acquisition of task-specific productions and modification of declarative representations in spatial-precuing tasks. Journal of Experimental Psychology: General, 117, 182–196. Reeve, T. G., & Proctor, R. W. (1990). The salient-features coding principle for spatial- and symbolic-compatibility effects. In R. W. Proctor & T. G. Reeve (Eds.), Stimulus–response compatibility (pp. 163–180). Amsterdam: North-Holland. Rosenbaum, D. A. (1980). Human movement initiation: Specification of arm, direction and extent. Journal of Experimental Psychology: General, 109, 444–474. Rosenbaum, D. A. (1983). The movement precuing technique: Assumptions, applications, and extensions. In R. A. Magill (Ed.), Memory and control of action (pp. 231–274). Amsterdam: North-Holland. Schlaghecken, F., & Eimer, M. (2004). Masked prime stimuli can bias ‘‘free” choices between response alternatives. Psychonomic Bulletin and Review, 11, 463–468. Stinear, C. M., Walker, K. S., & Byblow, W. D. (2001). Symmetric facilitation between motor cortices during contraction of ipsilateral hand muscles. Experimental Brain Research, 139, 101–105. Treisman, A. (1986). Features and objects in visual processing. Scientific American, 255, 114–125. Ugawa, Y., Hanajima, R., & Kanazawa, I. (1993). Interhemispheric facilitation of the hand area of the human motor cortex. Neuroscience Letters, 160, 153–155.
Contents lists available at ScienceDirect
Acta Psychologica journal homepage: www.elsevier.com/locate/actpsy
The influence of response grouping on free-choice decision making in a response selection task Michael A. Khan a,*, Stuart Mourton a, Eric Buckolz b, Jos J. Adam c, Amy E. Hayes a a
School of Sport, Health and Exercise Sciences, Bangor University, UK School of Kinesiology, University of Western Ontario, Canada c Department of Movement Sciences, Maastricht University, The Netherlands b
a r t i c l e
i n f o
Article history: Received 22 September 2009 Received in revised form 25 January 2010 Accepted 26 January 2010 Available online 25 February 2010 PsycINFO classification: 2330 2346 Keywords: Response cuing Free choice Response selection Response grouping
a b s t r a c t Previous research has demonstrated an advantage for the preparation of fingers on one hand over the preparation of fingers on two hands, and for the preparation of homologous fingers over that of nonhomologous fingers. In the present study, we extended the precuing effects observed with finger responses to response selection under free-choice conditions. Participants were required to choose from a range of possible responses following the presentation of a precue that indicated which response to prepare (go-to precue) or prevent (no-go-to precue). In Experiment 1 the choice was between homologous and non-homologous finger responses on the hand opposite to the precue while in Experiment 2 the choice was between finger responses on the same or different hand to the precue. In the go-to precue condition, the frequency of homologous finger choices was more frequent than non-homologous finger responses. Similarly, participants chose finger responses on the same hand as the precue regardless of whether they were instructed to prepare or prevent the precued response. The hand effect bias was stronger than the finger effect bias. These findings are consistent with the Grouping Model (Adam, Hommel, & Umilta, 2003). Ó 2010 Elsevier B.V. All rights reserved.
1. Introduction In order to interact with a complex environment, people will allocate attention selectively to events that they believe to be most pertinent. It is well known that reaction times to predicted stimuli are faster than to those that are unexpected (Posner, 1980). For example, a tennis player would react quicker to a shot to their forehand if it was expected than if they predicted a shot to their backhand. The influence of advance information on reaction time has numerous practical applications for strategy and tactics in sport and combat situations. From a theoretical standpoint, it also provides a basis for understanding the processes underlying selective attention and movement preparation. Interestingly, while most research on selective attention has been directed towards understanding how excitatory and inhibitory processes influence latency measures, relatively little attention has been given to how these processes influence which response is actually chosen. In real world situations (e.g., sport, combat and driving), people are often placed in conditions of free choice where they have a
* Corresponding author. Address: School of Sport, Health and Exercise Sciences, Bangor University, George Building, Bangor, Gwynedd, Wales LL57 2PZ, UK. Tel.: +44 (01248) 388275; fax: +44 (01248) 371053. E-mail address: [email protected] (M.A. Khan). 0001-6918/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.actpsy.2010.01.008
certain degree of freedom to use an internal mode of selection to choose from a range of possible responses. In the present research, we consider how advance information and the mechanisms underlying selective attention influence selection processes when participants have the freedom to choose a response option from several alternatives (i.e., free choice). A common technique used to investigate mechanisms of selective attention and movement preparation is the precue paradigm. Precues give participants advance information about features of an upcoming stimulus and/or required response. This advance information may be valid or invalid depending on whether stimulus/response features match that of the precue. The typical finding is that reaction time for valid precue trials is quicker compared to when no advance information is provided whereas there is a reaction time cost on invalid precue trials. These reaction time benefits and costs have served as a basis for testing theories of selective attention (e.g., Bekkering & Pratt, 2004; Egly, Driver, & Rafal, 1994) as well as processes underlying movement selection and preparation (e.g., Larish & Frekany, 1985; Rosenbaum, 1980). In most research on selective attention, either a many:1 or a 1:1 stimulus/response mapping has been employed. In the case of many:1 mappings, the response is specified from the outset and hence response selection processes are not involved. For 1:1 stimulus/response mappings, the response to be produced is fully
176
M.A. Khan et al. / Acta Psychologica 134 (2010) 175–181
specified by the characteristics of the stimulus. Therefore, regardless of whether one must react to a situation that is expected or unexpected, the selection of responses is externally determined since there is only one appropriate response on any particular trial. On the other hand, paradigms involving a 1:many stimulus/response mapping allow participants to make internal decisions when choosing freely from a subset of possible responses. It has been suggested that mechanisms of selection that underlie decision making in internal versus external choice situations are fundamentally different (e.g., Baylis, Tipper, & Houghton, 1997; Buckolz, Goldfarb, & Khan, 2004; Keller et al., 2006). However, it is interesting to note that most research on attention and speeded decision making has employed forced choice situations in which the selection of responses is externally determined. Although the study of how advance information influences speeded free choice is relatively limited, it has been demonstrated that the presentation of a subliminal cue influences response selection (Klapp & Haas, 2005; Schlaghecken & Eimer, 2004). In these studies, participants were first presented with a masked arrow that pointed left or right. This was followed by a stimulus arrow that pointed either to the left or right (forced choice trials) or a double headed arrow (free choice trials). It was demonstrated that on free choice trials, responses that corresponded to the precue were produced more often when the interval between the precue and the stimulus was short. At longer precue-stimulus intervals, there was a bias towards selecting responses that were opposite to the precue. Thus, it appears that the presentation of subliminal precues leads to activation and subsequent inhibition of motor responses resulting in systematic response selection biases. Posner, Rafal, Choate, & Vaughan (1985) have provided similar evidence that inhibition influences free choice by demonstrating that the selection of eye movements is biased away from locations that have been associated with earlier inhibition. The goal in the present research was to examine whether the instruction to prepare or prevent a precued response would influence which response option is chosen from a subsequently specified range of alternatives. The rationale behind our investigation was that when participants are instructed to prepare a particular response, the activation of this response will bias selection towards alternatives that share similar features (e.g., the hand or homologous finger). Conversely, when participants are instructed to prevent a particular course of action, the inhibition of a response option might bias response selection away from alternatives that share similar features. An underlying assumption is that response selection biases will arise from the grouping of response features between sets of alternatives. Strong evidence in support of this assumption stems from studies that have used the finger-cuing paradigm to investigate the efficiency of finger grouping processes. The finger-cuing paradigm was developed by Miller (1982), who adapted Rosenbaum’s movement precuing technique (Rosenbaum, 1980, 1983). In the finger-cuing task, a visual cue signal temporally precedes the target signal. The cue specifies a subset of two of four possible (keypress) responses (implemented by the index and middle fingers of both hands), thus prompting a process of subgroup making. In the hand-cued condition, the cue specifies two fingers on the same hand (e.g., the left-index finger and the left-middle finger). In the finger-cued condition, the cue specifies the homologous fingers on different hands (e.g., the two index fingers). In the neither-cued condition, the cue specifies non-homologous fingers on different hands (e.g., the left-middle and rightindex fingers). Also, a neutral (control) condition is included, which provides no advance information, and thus precludes the grouping and preparation of any combination of two finger responses. Cue effectiveness is inferred from a significant RT advantage for the informative cue conditions (i.e., hand-cued, finger-cued, and neither-cued) over the control, uninformative (neutral) cue condition.
The consistent finding from the finger-cuing paradigm is a pattern of differential cuing benefits that is apparent with short preparation intervals (i.e., intervals shorter than about 2–3 s). RTs are shortest for the hand-cued condition and longest for the neithercued condition, with RTs for the finger-cued condition being intermediate (for reviews see Adam et al., 2003, 2005;, Reeve & Proctor, 1990). A recent account of this pattern of differential cuing benefits is the Grouping Model (Adam et al., 2003, 2005), which is an extension of the salient-features coding principle advanced by Proctor & Reeve (1988) & Reeve & Proctor (1990). The key idea of the Grouping Model is that the individual elements of multi-element visual displays and multi-element response arrays are not processed independently but are pre-attentively organized or ‘‘grouped” according to low-level grouping factors that depend on stimulus driven factors (e.g., Gestalt principles) and on response-related factors (e.g., inter-response linkages). Preattentive processing is done quickly, effortlessly and in parallel without the need of focused attention (cf. Treisman, 1986). In other words, the basic assumption is that, each stimulus set and each response set has a default organization established automatically by the bottom-up computation of perceptual and motoric units or subgroups; this process is fast and effortless. With additional, top-down processing, however, alternative organizations can be attained; this process is slow and effortful. Thus, the pattern of cuing effects that emerge in the finger-cuing task critically depends on the nature of these default groupings and on the time available to reorganize these representations, if necessary. According to the Grouping Model, the processing advantage of hand-cues simply reflects the natural, strong grouping of two fingers on the same hand. The co-activation of directly adjacent and overlapping cortical finger representations in areas of the motor cortex corresponding to the same-hand finger set could be at the basis of this grouping (e.g., Dechent & Frahm, 2003). The bilateral finger- and neither-cues, on the other hand, are more difficult to process because they require slow, effortful, top-down modulation to breakup the anatomically based left–right motor organization and to create a new motor organization represented in two hemispheres. Moreover, the advantage of finger-cues over neither-cues can be attributed to the fact that finger-cues require the grouping of homologous fingers whereas neither-cues require the grouping of two different, non-homologous fingers. Because homologous fingers are neurally and functionally linked (e.g., Stinear, Walker, & Byblow, 2001; Ugawa, Hanajima, & Kanazawa, 1993), they are easier to group or co-activate than non-homologous fingers. Hence, the existence of facilitatory connections between homologous motor areas in the brain implies that the preparation and activation of one finger may spill over to its homologous counterpart in the opposite hemisphere. In the present experiments, we extended the precuing effects observed with finger responses to response selection under freechoice conditions. A novel paradigm was employed in which participants were required to choose from a range of possible responses following the presentation of advance information. On each trial, a precue was presented that indicated to the participant which response to prepare (i.e., go-to precue). This was followed by the imperative stimulus that specified a subset of possible responses rather than a specific response (1: many stimulus–response mapping). In Experiment 1, the imperative stimulus specified the finger responses on either the left or right hand. When the precued response was on the same hand specified by the imperative stimulus, participants were required to produce the precued response. However, if the imperative stimulus specified the other hand, participants were placed in a free-choice situation. Hence, they could choose between a homologous and nonhomologous finger relative to the finger specified by the precue. Based on the above reviewed evidence that supports homologous
M.A. Khan et al. / Acta Psychologica 134 (2010) 175–181
finger grouping over non-homologous finger grouping, we expected participants in a free-choice situation to opt more often for a homologous than for a non-homologous finger response. We also employed trials in which participants were instructed to prevent the precued response (no-go-to precue). In these cases, it was expected that the inhibition of the precued response would bias choices away from a homologous finger response in the free-choice situation. 2. Experiment 1 2.1. Method 2.1.1. Participants Twenty-four self declared, right hand dominant, university students served as participants in the study (12 males, 12 females, ages 18–35 years). All participants in this Experiment and Experiment 2 were naive to the hypothesis being tested and inexperienced at the experimental task. They gave their informed consent prior to participation in the study. The experimental protocols were approved by the Ethics Committee of the School of Sport, Health and Exercise Sciences, Bangor University for research involving human participants in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki. 2.1.2. Apparatus Participants sat in a chair in front of a computer monitor that was placed on a table top. The monitor was approximately 45 cm from the eyes of the participants. The display on the monitor consisted of four squares (4 4 cm) arranged horizontally with a separation of 4 cm (side to side) on a black background. The squares had a white outline and were not filled. The middle and index fingers of the left hand were placed on the ‘‘F’ and ‘‘G” keys, respectively, while the index and middle fingers of the right hand were placed on the ‘‘J” and ‘‘K” keys, respectively. The four squares on the monitor were mapped in a spatially compatible manner with the four finger presses such that the leftmost square corresponded to the left-middle finger keypress, and so on. 2.1.3. Procedure At the beginning of each trial, the four target squares appeared on the screen. After a delay of 2000 ms a precue was presented by coloring green the outlines of one of the squares for a duration of 1000 ms. Then, following an interval of either 100 or 750 ms, the imperative stimulus was presented between the second and third squares. Thus, there were two stimulus-onset asynchronies (SOAs) between precue and imperative stimulus onset, namely 1100 and 1750 ms. The imperative stimulus consisted of an arrow pointing to the right or left (see Fig. 1). The imperative stimulus arrow defined a subset of two responses. An arrow pointing to the right specified the two right hand finger responses while an arrow pointing to the left corresponded to the two left hand responses.
Precue
Stimulus Forced-choice
177
The imperative stimulus remained on screen until the participant made a response. In one block of trials (i.e., go-to precue), participants were instructed to press the response corresponding to the precue if the stimulus arrow pointed to the same side as the precue. This was a forced choice situation in that only one alternative was appropriate. For example, if the right-index finger was precued and the imperative stimulus arrow pointed to the right, participants had to depress the right-index finger (i.e., J key). However, if the arrow pointed to the opposite side to that of the precue, the participant now had a free-choice situation in that they could choose either of the two responses on the left hand. In another block of trials (no-go-to precue), participants were told to prevent making the response that corresponded to the precue. Therefore, if the stimulus arrow pointed to the same side as the precue, participants were required to produce the response on that hand that did not correspond to the precue. For example, if the right-index finger was precued and the arrow pointed to the right, participants were required to respond with their right middle finger. If the arrow pointed to the left, participants again had a free-choice situation and could choose between the index and middle fingers of the left hand. Each participant performed two blocks of trials, one in which the precues were designated as go-to precues and one in which precues were designated as no-go-to precues. For each block of trials, the eight combinations of the four precue positions and two imperative stimulus directions were randomized in a pseudo random fashion such that each combination was administered before any was repeated. Each combination was administered twenty times giving a total of 160 trials per block. Since each combination of precue position and stimulus direction occurred equally often, there was a probability of 50% that the imperative stimulus would point to the same or opposite side to that of the precue. The order of the go-to and no-go-to precue trial blocks was counterbalanced between participants. 2.1.4. Analytic methods Of primary interest in the present experiment were those trials in which the stimulus arrow pointed to the opposite direction to that of the precue (i.e., free choice trials). Specifically, we were interested in whether participants would choose the finger response that was homologous or non-homologous to the precue. For the free choice data, since the homologous finger condition and the non-homologous finger condition are interdependent, they cannot be analysed as independent levels of a Choice factor. Hence, to determine whether one finger response was produced more often than the other, the difference in response frequencies was calculated between the homologous finger and non-homologous finger values for both the go-to and no-go-to conditions. One-sampled t-tests were then used to test whether this difference was significantly greater than zero. A preliminary analysis revealed that these difference scores were not affected by SOA for either Precue Condition, and therefore the SOA factor will not be discussed further. Forced choice and free choice RTs were analysed using a 2 Precue Condition (go-to, no-go-to) 3 Choice (forced choice, homologous finger, non-homologous finger) fully repeated measures ANOVA. 2.2. Results and discussion
Free-choice LM
LI
RI
RM
Fig. 1. Sequence of events on each trial from precue to imperative stimulus presentation on forced and free choice trials in Experiment 1 (LM – left middle; Li – left index; RI – right index; RM – right middle). All squares were outlined squares, but are depicted as filled in for clarity.
2.2.1. Errors The overall error rate in the go-to trial blocks was 9.6% (see Table 1). On forced choice trials, the most frequent errors were those in which participants produced the incorrect finger response on the same hand as the precue. The most frequent errors on the free choice trials were those in which the precued response was
178
M.A. Khan et al. / Acta Psychologica 134 (2010) 175–181
Table 1 Percentage of errors for the go-to and no-go-to precues under the free and force choice conditions in Experiment 1. Errors are coded in relation to the precued finger (SH = same hand; DH = different hand; SF = same finger; DF = different finger). Go-to precue
No-go-to precue
Forced choice
Forced choice
SH/DF
DH/SF
DH/DF
5.6
0.6
0.1
Free choice
Precue
DH/SF
DH/DF
4.5
0.6
1.1
Free choice
Precue
SH/DF
Precue
SH/DF
2.9
0.4
0.4
1.4
produced. The overall error rate in the no-go-to trial blocks was 8.0%. The highest error rates were on forced choice trials in which the precued response was produced. Trials in which errors were committed were excluded from the following response frequency and reaction time analyses. 2.2.2. Response frequencies The proportion of free choice trials for which participants produced homologous and non-homologous finger responses to that of the precue are shown in Fig. 2. One-sampled t-tests revealed that for the go-to condition, the difference was greater than zero (t(23) = 3.50; p = 0.002), indicating that homologous finger responses were significantly more frequent than the non-homologous finger responses. Previous research has revealed that RTs were quicker when homologous finger pairs were precued compared to when nonhomologous fingers were precued (Adam et al., 2003, 2005). In those studies, two responses were specified by the precue while the imperative stimulus specified one of the two precued responses. The results of the current study in which participants were given a free choice following the presentation of a single response precue are consistent with these findings. The finding that participants chose the finger response that was homologous to the precue more often than the non-homologous finger indicates that the preparation of a single response does bias decision making towards alternatives that share similar features (i.e., homologous finger). For the no-go-to condition, the difference between the homologous and non-homologous finger responses was not significantly different from zero (t(23) = 0.26; p = 0.8). We had expected that the instruction to prevent a particular response would bias deci-
homologous non-homologous
0.8
Response Frequency
2.2.3. Reaction times The analysis of RT revealed a significant Precue Condition Choice interaction, F(2, 46) = 66.2, p < 0.01. Breakdown of this interaction using Tukey HSD post hoc tests (p < 0.05) revealed that in the go-to condition, RTs were shorter for the forced choice compared to free choice trials, a finding that reflects the instruction to prepare the precued response (see Table 2). Also, on free choice trials, RTs were shorter for homologous finger than for non-homologous finger responses. Hence, not only did participants choose homologous finger responses more often, but RTs were shorter for these responses compared to non-homologous fingers. In the no-go-to condition, RTs were longer in the forced choice condition compared to when a non-homologous response was chosen on free choice trials. This may reflect a strategy adopted by participants to avoid using the hand that contained the no-go-to precue. Similar to the response frequency results in the no-go-to condition, there was no difference in RT between the homologous and non-homologous responses on free choice trials. 3. Experiment 2 In Experiment 1, the imperative stimulus specified either the two left or right hand responses. Hence, on free choice trials, the choice was between homologous and non-homologous finger responses relative to the precue. In Experiment 2, the imperative stimulus specified the two index fingers or the two middle fingers. In free choice trials, the choice was between (non-homologous) finger responses located on the same or different hand relative to the finger specified by the cue. Based on evidence that within-hand grouping is much stronger than between-hands grouping, we expected participants in the free-choice situation to show a robust selection preference for the same-hand finger response. Furthermore, this selection bias for a same-hand over a different-hand response alternative should be stronger than the selection bias demonstrated in Experiment 1 for a homologous over a nonhomologous response alternative. This would reflect the dominance of hand grouping over (homologous) finger grouping. 3.1. Method 3.1.1. Participants Twenty-four self declared, right hand dominant, university students served as participants in the study (12 males, 12 females, ages 18–35 years).
1 0.9
sion making away from responses that shared similar features. However, this was not the case since the frequency of non-homologous finger responses was not significantly different from homologous responses.
3.1.2. Apparatus Same as in Experiment 1.
0.7 0.6
3.1.3. Procedure The procedures were similar to Experiment 1 except that the imperative stimulus now consisted of arrows pointing inwards or
0.5 0.4 0.3 0.2
Table 2 Mean (standard deviation) reaction times (ms) under go-to and no-go-to precue conditions in Experiment 1 for forced choice and free-choice conditions.
0.1 0 go-to
no-go-to
Precue Condition Fig. 2. Response frequencies for homologous and non-homologous finger options for the go-to and no-go-to precues in the free-choice condition of Experiment 1.
Response
Go-to precue
No-go-to precue
Forced choice Free choice: homologous finger Free choice: non-homologous finger
424 (67) 477 (64) 490 (69)
482 (69) 475 (50) 467 (53)
179
M.A. Khan et al. / Acta Psychologica 134 (2010) 175–181
3.1.4. Analytic methods The statistical analyses were analogous to those of Experiment 1. As in Experiment 1, initial analyses indicated no significant effects of SOA on response frequencies, and hence this factor will not be considered further.
3.2. Results and discussion 3.2.1. Errors The overall error rate in the go-to trial blocks was 8.8% (see Table 3). On forced choice trials, the highest errors rates occurred when participants produced the homologous finger response on the opposite hand to the precue. On free choice trials, the most frequent errors were those in which the precued finger response was produced. The overall error rate in the no-go-to trial blocks was 8.9%. On forced choice trials, the highest errors occurred when participants produced the precued response. On free choice trials, the most frequent errors were those in which the homologous finger response on the opposite hand to the precue was produced.
Precue
Stimulus
><
Forced-choice
<>
Free-choice
Table 3 Percentage of errors for the go-to and no-go-to precues under the free and force choice conditions in Experiment 2. Errors are coded in relation to the precued finger (SF = same finger; DF = different finger; SH = same hand; DH = different hand). Go-to precue
No-go-to precue
Forced choice
Forced choice
SF/DH
DF/SH
DF/DH
Precue
DFSH
DF/DH
2.8
0.8
0.1
2.7
0.4
1.9
Free choice
Free choice
Precue
SF/DH
Precue
SF/DH
4.9
0.2
0.1
3.0
Trials in which errors were committed were excluded from the following response frequency and reaction time analyses. 3.2.2. Response frequencies Of primary interest in the present experiment was whether participants would choose the finger response on the same or other hand to that of the precue on free choice trials. The proportion of free choice trials for which the participants produced a finger response on the same and different hand to that of the precue are shown in Fig. 4. Similar to Experiment 1, a difference score was calculated between the same and different hand frequencies for both the go-to and no-go-to conditions. One-sampled t-tests revealed that the difference was greater than zero for both the go-to condition (t(23) = 14.30; p < 0.001) and the no-go-to condition (t(23) = 10.41; p < 0.001). As shown in Fig. 4, responses on the same hand to that of the precue were chosen more often regardless of whether participants were instructed to prepare or prevent the precued response. A paired-samples t-test revealed that the tendency to select the same-hand finger response was significantly greater in the go-to condition than in the no-go-to condition (t(23) = 3.57; p = 0.002). This relatively large bias towards selecting a response from the same hand as the precue reflects the strong nature of within-hand grouping. Previous research has demonstrated that RTs were shortest when two responses on the same hand were precued compared to when two homologous or nonhomologous fingers on different hands were precued. The present results imply that this within-hand advantage also emerges in the form of response biases under free-choice conditions. 3.2.3. Reaction times As mentioned above, there was a strong tendency to choose responses on the same hand as the precue. Data from one participant
1
same hand different hand
0.9
Response Frequency
outwards (see Fig. 3). Arrows pointing inwards specified the index fingers of each hand while arrows pointing outwards specified the middle fingers. Therefore, in contrast to Experiment 1 where the imperative stimulus specified finger responses on the same hand, the imperative stimulus now specified homologous fingers on separate hands. Hence, in free-choice situations, the choice was in effect between hands instead of between fingers. Participants performed one block of trials under the go-to precue condition and another block under the no-go-to precue condition. As in Experiment 1, participants were instructed to prepare the precued response in the go-to condition. For example, if the right-index finger was precued and the stimulus arrows pointed inwards, participants were required to depress the right-index finger. However, if the arrows pointed outwards, participants could choose either the right or left-middle finger. In the no-go-to condition, participants were instructed to prevent making the response indicated by the precue. Hence, if the arrows pointed to a subset of finger responses that included the precued finger response, participants were required to produce the response opposite to the position of the precue. For instance, if the right-index finger was precued and the arrows pointed inwards, participants were required to respond with the left-index finger. However, if the arrows pointed outwards, participants were again free to choose between the middle finger of the left hand and the middle finger of the right hand. For each block of trials, the eight combinations of the four precue positions and two stimulus arrows were randomized in a pseudo random fashion such that each combination was administered before any was repeated. Each combination was administered twenty times giving a total of 160 trials per block. The order of the go-to and no-go-to precue trial blocks was counterbalanced between participants.
0.8 0.7 0.6 0.5 0.4 0.3 0.2
LM
LI
RI
RM
0.1 0
Fig. 3. Sequence of events on each trial from precue to stimulus presentation on forced and free choice trials in Experiment 2 (LM – left middle; Li – left index; RI – right index; RM – right middle). Stimuli were outlined squares, but are depicted as filled in for clarity.
go-to
no-go-to
Precue Condition Fig. 4. Response frequencies for same hand and different hand options for the go-to and no-go-to precues in the free-choice condition of Experiment 2.
180
M.A. Khan et al. / Acta Psychologica 134 (2010) 175–181
were excluded from the RT analysis because there were no different-hand responses in at least one of the experimental conditions. The analysis of RTs from the remaining 23 participants was performed using a 2 Precue Condition (go-to, no-go-to) 3 Choice (forced choice, same hand, different hand) ANOVA. This analysis revealed a significant Precue Condition Choice interaction, F(2, 44) = 39.6, p < 0.001. Breakdown of this interaction using Tukey HSD post hoc tests (p < .05) revealed that RTs in the go-to precue condition were faster for the forced compared to free choice trials (see Table 4). Also, when participants chose a response that was on the same hand as the precue, RTs were faster than when they chose a response on the other hand. This again demonstrates the strong influence of within-hand grouping on response latencies. There were no differences in RTs in the no-go-to condition. 3.2.4. Between experiment comparison In order to compare the magnitude of free choice biases between Experiment 1 and Experiment 2, the difference in frequencies between same and different finger responses in Experiment 1 and between same and different hand responses in Experiment 2 were submitted to a 2 Experiment 2 Precue (go-to, no-go-to) ANOVA with repeated measures on the second factor. A main effect of Experiment, F(1, 46) = 136, p < 0.001, indicated that the difference between same and different-hand response frequencies in Experiment 2 was greater than the difference between homologous and non-homologous finger frequencies in Experiment 1. Also, a main effect of Precue, F(1, 46) = 21.9, p < 0.001, indicated that response biases were greater in the go-to compared to nogo-to condition. These results further demonstrate stronger within compared to between hand grouping of finger responses. 4. General discussion Past research has shown that when two of four finger responses are precued, RTs are fastest when the two precued responses are on the same hand, slowest when the precue specifies non-homologous fingers from each hand, and intermediate RTs when homologous fingers are precued (Adam & Pratt, 2004). In the present study, we extended these precuing effects to free-choice situations. Our primary goal was to investigate if a precued response would bias the subsequent response selection processes. The results of Experiment 1 revealed that for go-to precues, participants selected the homologous finger responses in free-choice conditions more often than non-homologous responses. No-go-to precues did not have an influence on response selection. In Experiment 2, participants were given a choice between homologous finger responses on separate hands. There was a strong bias towards choosing responses on the same hand as the precue. Although this was the case for both go-to and no-go-to precues, the bias towards within-hand choices was stronger in the go-to condition. In order to account for the effects of advance information on RTs of finger responses, Adam et al. (2003, 2005) have proposed a Grouping Model that takes into account coding processes at both stimulus and response levels. According to the Grouping Model, arrays of stimuli and responses are separately organised pre-attentively into subgroups. In cases in which the precue defines a
Table 4 Mean (standard deviation) reaction times (ms) under go-to and no-go-to precue conditions in Experiment 2 for forced choice and free-choice conditions. Response
Go-to precue
No-go-to precue
Forced choice Free choice: same hand Free choice: different hand
475 (38) 547 (54) 574 (81)
536 (52) 542 (48) 548 (56)
subgroup of stimuli that maps directly to a subgroup of responses, a fast automatic selection of responses occurs. When the precue does not correspond to a subgroup of stimuli or there is not a direct relation between stimulus and response subgroups, effortful processing is needed to select and activate responses. The Grouping Model has been based predominantly on experiments in which the precue specifies a subset of responses (e.g., two of four finger keypresses) from which one would be specified by the imperative stimulus. Hence, the efficiency of subgroup formation is contingent upon which responses are specified by the precue. The paradigm in the present study differs from past work in that the precue specified only one response (not two). Participants were then required to choose from a subset of responses that was specified by the imperative stimulus. Hence, in contrast to the classic finger-precuing paradigm, the precue itself did not specify a subgroup of responses. Of interest was whether the preparation of a single response would lead to subgroup formation by co-activation of other responses that shared similar features with the precued response. If this were the case, response selection would be biased towards responses that were within the same subgroup as the precued response. More specifically, based on the assumption that fingers on the same hand form a strong, well-defined subgroup, the activation of one finger response would result in simultaneous activation of other fingers on the same hand. This increased activation would bias response selection in free-choice situations towards the more highly activated responses on the same hand as the precue compared to responses on the other hand. The findings of Experiment 2 substantiated this assumption by showing that participants strongly preferred a response option on the same hand as the precue compared to a response option on the other hand. Similarly, the activation of a single finger response may not only be limited to responses on the same hand but also to fingers on the other hand that share a common feature. This was demonstrated in Experiment 1 where response selection was biased towards homologous finger pairings on the opposite hand to that of the precue. Since both experiments revealed that the instruction to prepare a single response influenced free choice response selection, it appears that both within-hand and homologous finger pairings belong to well-defined (motor) subgroups that are neurally linked (e.g., Dechent & Frahm, 2003; Stinear et al., 2001). It is possible that participants may have adopted a strategic approach to subgroup formation upon presentation of the precue in order to reduce the number of alternatives. For example, in Experiment 1, when the right-index finger was precued, participants may have prepared the right-index finger (in the event of a forced choice stimulus) along with the left-index finger (in the event of a free choice stimulus). Similarly, in Experiment 2, when the rightindex finger was precued, participants may have simultaneously prepared the right middle finger. However, this does not appear to be the case since reaction times for the precued responses were clearly shorter than other responses. Hence, participants were preparing the precued response to a level beyond all other alternatives. In Experiment 1, participants were required to choose between homologous or non-homologous finger responses on the opposite hand to the precue. In Experiment 2, the choice was between finger responses on the same or opposite hand to the precue. Comparison between experiments indicated that fingers on the same hand constituted a stronger subgroup than homologous finger pairings on different hands. When participants were required to choose between finger responses on different hands, the tendency to choose a response on the same hand as the precue was stronger than choosing homologous finger responses on the hand that did not contain the precued response. These findings then are compatible with the Grouping Model since they conform to the pattern of
M.A. Khan et al. / Acta Psychologica 134 (2010) 175–181
differential precuing benefits where hand-cues are most effective, neither-cues (non-homologous fingers) least, and finger-cues (homologous fingers) intermediate, indicating that within-hemisphere co-activation of finger responses is stronger than between-hemisphere co-activation of finger responses. Although go-to precues had an influence on response frequencies and reaction times, the influence of no-go-to precues on response selection was relatively weak. Given the evidence for inhibitory connections within and between the two hemispheres (e.g., Ferbert et al., 1992), we had expected that the instruction to prevent a particular response would lead to the inhibition of other responses that shared similar features (i.e., responses within the same subgroup). However, the results of Experiment 1 did not reveal a bias away from homologous finger responses. In Experiment 2, on the other hand, there was a significant predisposition to choose responses on the same hand as the no-go-to precue but this bias was not as strong when compared to go-to precues. Hence, these findings do not unequivocally support our hypothesis that the prevention of a particular response biases selection away from other responses in the same subgroup. This would be in contrast to other studies that have shown that responses are biased away from the direction of a subliminal precue (i.e., masked primes) at long precue-stimulus ISIs (e.g., Schlaghecken & Eimer, 2004). It was proposed that the initial activation of the cued response was followed by inhibition that consequently led to a reduced frequency and increased RT for the cued response. However, one important difference in the design of our study was that the no-go-to response could not be used whereas it remained a viable option in past work. Hence, it may be that while the inhibitory effects associated with a cued response bias response selection away from that response, they are not of sufficient strength to transfer to other responses. It is also possible that the instruction to prevent a particular response was not a true test of how its inhibition influences other responses. This is because participants may have prepared alternative responses upon presentation of a no-go-to precue; such a recoding strategy would be akin to ‘‘indirect cuing” making it similar to the go-to condition. Hence, the results of the no-go-to trials should be interpreted with caution at this point. In summary, the present results offer strong support for the Grouping Model proposed by Adam et al. (2003, 2005). Past work has clearly demonstrated the benefit of subgroup formation on response preparation through the shortening of response latencies. In the present study, we extended this to free-choice situations. When participants were instructed to prepare a response, response selection under free-choice conditions was biased towards responses that shared similar features with the precued response. This finding is a powerful demonstration of subgroup formation (or response co-activation) even in cases in which only one response is specified by the precue.
181
References Adam, J. J., Hommel, B., & Umilta, C. (2003). Preparing for perception and action (I): The role of grouping in the response-cuing task. Cognitive Psychology, 46, 302–358. Adam, J. J., Hommel, B., & Umilta, C. (2005). Preparing for perception and action (II): Automatic and effortful processes in response cueing. Visual Cognition, 12, 1444–1473. Adam, J. J., & Pratt, J. (2004). Dissociating visual attention and effector selection in spatial precuing tasks. Journal of Experimental Psychology: Human Perception and Performance, 30, 1092–1106. Baylis, G. C., Tipper, S. P., & Houghton, G. (1997). Externally cued and internally generated selection: Differences in distractor analysis and inhibition. Journal of Experimental Psychology: Human Perception & Performance, 23, 1617–1630. Bekkering, H., & Pratt, J. (2004). Object-based processes in the planning of goaldirected hand movements. The Quarterly Journal of Experimental Psychology, 8, 1345–1368. Buckolz, E., Goldfarb, A., & Khan, M. A. (2004). The use of distractor assigned response slows later responding in a location negative priming task. Perception and Psychophysics, 66, 837–845. Dechent, P., & Frahm, J. (2003). Functional somatotopy of finger representations in human primary motor cortex. Human Brain Mapping, 18, 272–283. Egly, R., Driver, J., & Rafal, R. D. (1994). Shifting visual attention between objects and locations: Evidence from normal and parietal lesion subjects. Journal of Experimental Psychology: General, 123, 161–177. Ferbert, A., Priori, A., Rothwell, J. C., Day, B. L., Colebatch, J. G., & Marsden, C. D. (1992). Interhemispheric inhibition of the human motor cortex. Journal of Physiology, 453, 525–546. Keller, P. E., Wascher, E., Prinz, W., Waszak, F., Koch, I., & Rosenbaum, D. A. (2006). Differences between intention-based and stimulus-based actions. Journal of Psychophysiology, 20, 9–20. Klapp, S. T., & Haas, B. W. (2005). Nonconscious influence of masked stimuli on response selection is limited to concrete stimulus–response associations. Journal of Experimental Psychology: Human Perception and Performance, 31, 193–209. Larish, D. D., & Frekany, G. A. (1985). Planning and preparing expected and unexpected movements: Reexamining the relationships of arm, direction, and extent of movement. Journal of Motor Behavior, 17, 168–189. Miller, J. (1982). Discrete versus continuous models of human information processing: In search of partial output. Journal of Experimental Psychology: Human Perception and Performance, 8, 273–296. Posner, M. I. (1980). Orienting of attention. The VIIth Sir Frederic Bartlet Lecture. Quarterly Journal of Experimental Psychology, 32, 3–25. Posner, M. I., Rafal, R. D., Choate, L. S., & Vaughan, J. (1985). Inhibition of return: Neural basis and function. Cognitive Neuropsychology, 2, 211–228. Proctor, R. W., & Reeve, T. G. (1988). The acquisition of task-specific productions and modification of declarative representations in spatial-precuing tasks. Journal of Experimental Psychology: General, 117, 182–196. Reeve, T. G., & Proctor, R. W. (1990). The salient-features coding principle for spatial- and symbolic-compatibility effects. In R. W. Proctor & T. G. Reeve (Eds.), Stimulus–response compatibility (pp. 163–180). Amsterdam: North-Holland. Rosenbaum, D. A. (1980). Human movement initiation: Specification of arm, direction and extent. Journal of Experimental Psychology: General, 109, 444–474. Rosenbaum, D. A. (1983). The movement precuing technique: Assumptions, applications, and extensions. In R. A. Magill (Ed.), Memory and control of action (pp. 231–274). Amsterdam: North-Holland. Schlaghecken, F., & Eimer, M. (2004). Masked prime stimuli can bias ‘‘free” choices between response alternatives. Psychonomic Bulletin and Review, 11, 463–468. Stinear, C. M., Walker, K. S., & Byblow, W. D. (2001). Symmetric facilitation between motor cortices during contraction of ipsilateral hand muscles. Experimental Brain Research, 139, 101–105. Treisman, A. (1986). Features and objects in visual processing. Scientific American, 255, 114–125. Ugawa, Y., Hanajima, R., & Kanazawa, I. (1993). Interhemispheric facilitation of the hand area of the human motor cortex. Neuroscience Letters, 160, 153–155.