Journal of Experimental Social Psychology 42 (2006) 479–490 www.elsevier.com/locate/jesp
Stopping dead in oneÕs tracks: Motor inhibition following incidental evaluations Benjamin M. Wilkowski *, Michael D. Robinson Psychology Department, North Dakota State University, Fargo, ND 58105, USA Received 31 July 2004; revised 3 August 2005 Available online 7 November 2005
Abstract Motivational theories of affect contend that negative stimuli should slow motor behavior. However, this claim has been supported mainly by animal research. The current investigation establishes, perhaps for the first time, that incidental negative primes slow subsequent motor behavior among humans. In addition to examining choice reaction time (Studies 1 and 2), we introduce a method for examining the speed of continuous motor behavior (Study 3). In all studies, we found that behavior was slowed following, or in the context of, negative stimuli. It is concluded that, in certain situations, the mere presence of negative stimuli inhibits the speed of subsequent motor behavior. 2005 Elsevier Inc. All rights reserved. Keywords: Affect; Freezing; Motivation; Automatic evaluation; Inhibition
Introduction In speculating on the likely function of affect and emotion, theorists have raised a number of possibilities (see Ekman & Davidson, 1994; chapter 3). However, one answer that has emerged again and again is that affect must serve a useful purpose in guiding individuals toward adaptive behaviors (e.g., Lang, 1995; Neumann, Fo¨rster, & Strack, 2003). In particular, affect should help guide individuals towards rewards and away from punishments (Cacioppo, Gardner, & Bernston, 1999; Higgins, 1997; Watson, Wiese, Vaidya, & Tellegen, 1999). Such considerations suggest that affect and action tendencies should be closely associated. Drawing on animal research, a number of theorists have suggested that, under certain conditions, individuals may ‘‘freeze’’ or slow movement following the detection of negative stimuli (Fanselow, 1994; Gray, 1987; Lang, Bradley, & Cuthbert, 1997). Evolutionarily speaking, this behavior
*
Corresponding author. Fax: +1 701 231 8426. E-mail address:
[email protected] (B.M. Wilkowski).
0022-1031/$ - see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.jesp.2005.08.007
probably first served the function of avoiding detection by predators. Likewise, theorists have speculated that the presence of positive stimuli serve as signals to safety and opportunity, and thus facilitate or energize on-going motor behavior (Depue & Lenzenweger, 2001; Gray, 1987). The current research is designed to test these speculations behaviorally among humans. Motivational theories of affect How does affect guide individuals to approach rewards and avoid punishments? A theoretical consensus is building that two separate motivational systems serve these functions independently of one another (e.g., Carver, Sutton, & Scheier, 2000; Lang, 1995; Watson et al., 1999). Specifically, one system is responsible for the detection and attainment of rewarding or positive stimuli. A separate system is responsible for the detection and avoidance of punishing or negative stimuli. Such models have emerged from a diverse set of research areas, including animal conditioning (e.g., Crawford & Masterson, 1982), animal neuropsychology (e.g., Depue & Lenzenweger, 2001; Gray, 1987), human neuropsychology (e.g., Cacioppo et al., 1999; Davidson,
480
B.M. Wilkowski, M.D. Robinson / Journal of Experimental Social Psychology 42 (2006) 479–490
1992), self-reported mood (e.g., Tellegen, 1985; Watson, 2000), and psychophysiology (e.g., Fowles, 1980; Lang et al., 1997). According to such models, these motivational systems detect rewarding or punishing stimuli, and subsequently trigger appropriate changes in subjective feeling states (Watson, 2000; Watson et al., 1999), physiological responses (Fowles, 1980; Lang et al., 1997), and behavioral action tendencies (Fanselow, 1994; Gray, 1987; Lang et al., 1997). What is the nature of these behavioral action tendencies? Pavlov (1927) first documented that one consequence of exposing an animal (e.g., a rat) to a negative stimulus (e.g., a predator) is freezing behavior. Freezing behavior consists of slowing motor movement, as if the animal is trying to avoid detection by a predator. However, punishing stimuli can also produce a number of other overt behaviors, such as flight or fight. Fanselow (1994) reviewed evidence from the animal psychology literature, and concluded that the nature of the automatic affective motor response is contingent upon the nature of the situation. When a rat first detects a predator, but believes that it has not been detected by the predator, the rat seeks to avoid detection. It does so by freezing or slowing all motor movement. However, when the rat believes that it has been detected by the predator, the goal is now to avoid the predator in a more active manner. It does so by flight if possible and fight if necessary. Fanselow thus predicts that fight/ flight responses will be primed when responding directly to a negative stimulus; however freezing responses will be primed when one is behaving within the context of a negative stimulus, but not responding directly to that stimulus. Although a number of theories (e.g., Gray, 1987; Lang et al., 1997) have used animal research on freezing as a model for the nature of human motivational systems, the corresponding behavioral research among human populations is at this point, incomplete. As such, generalizations to the nature of human affect are currently uncertain. However, a great deal of social psychological research on automatic evaluation is consistent with FanselowÕs theory, and thus provides a useful framework for exploring predictions related to freezing behavior. Automatic evaluation One assumption implicit in FanselowÕs (1994) theory is that individuals will automatically evaluate stimuli within their environment. A plethora of studies within the social psychology literature on attitudes demonstrate that this is indeed the case. Fazio, Sanbonmatsu, Powell, and Kardes (1986) first demonstrated automatic evaluation by presenting participants with a to-be-ignored prime, followed by a target word to be evaluated. Positive targets were evaluated faster following positive primes, whereas negative targets were evaluated faster following negative primes, suggesting that the prime had activated the cognitive substrates of evaluation within memory (see Klauer & Musch, 2003, for a review).
Subsequent studies have linked the automatic evaluation effect to FanselowÕs (1994) work more directly. Specifically, these studies have examined the extent to which affective primes bias behavioral responses (for a review, see Neumann et al., 2003). For example, authors have shown that people are quicker to respond to a positive stimulus with an approach (relative to avoidance) behavior, whereas they are quicker to respond to a negative stimulus with an avoidance (relative to approach) behavior (see Chen & Bargh, 1999; Markman & Brendl, 2005; Neumann et al., 2003). Importantly, such studies invariably required participants to respond directly to the negative stimulus itself, and thus correspond to FanselowÕs (1994) predictions that fight or flight behaviors are primed when responding directly to a negative stimulus. However, tests of FanselowÕs (1994) prediction that motor inhibition (rather than avoidance behaviors) should result when one is only in the context of negative stimuli are quite sparse with regard to human participants. We could locate only one study examining automatic motor inhibition effects among humans. Derryberry (1991) used a rather complex design to separate the attentional, motoric, and arousal-related effects of affect, and found some support for the idea that negative affect slows subsequent motor behavior. However, given the complexities of his design (which would require more elaboration than is useful here), we sought to establish such affective priming effects in a more straightforward and simple manner. We hypothesized, following Fanselow, that motor inhibition should result when one acts within the context of a negative stimulus, but the action is not contingent on the negative stimulus itself. Likewise, we tested the notion that motor facilitation should result when one acts within the context of a positive stimulus (see Depue & Collins, 1999; Gray, 1987). How automatic is automatic evaluation? As research on automatic evaluation has progressed, it has become apparent that evaluative processes are automatic with respect to certain criteria, but not others (Klauer & Musch, 2003). Specifically, it appears that evaluations are often automatic with respect to the criterion of unintentionality (see Bargh, 1989, for a helpful overview of the criteria of automaticity). That is, people exhibit evidence of evaluating prime stimuli despite no intention to do so (see Wentura & Rothermund, 2003). However, research has also demonstrated that evaluations are often not automatic with respect to the criterion of efficiency. That is, people often do not exhibit evidence of evaluating primes that are outside the focus of attention (Musch & Klauer, 2001; Pessoa, McKenna, Gutierrez, & Ungerleider, 2002). In light of such considerations, we sought to demonstrate that motor inhibition and activation are unintentional in nature. That is, we gave participants no goal of evaluating the stimuli in these tasks. However, we made
B.M. Wilkowski, M.D. Robinson / Journal of Experimental Social Psychology 42 (2006) 479–490
no attempt to demonstrate that these effects are efficient in nature. Indeed, we believe it likely that the effects reported here required focused attention. Thus participants in the current studies were encouraged to process affective primes, although they were given no goal of evaluation. Response selection and response execution Recent findings in cognitive psychology provide a useful context for examining motor behavior. It is typical to view the selection of a response as the endpoint of cognitive operations (Grice, Nullmeyer, & Spiker, 1982; McClelland, 1979; Sternberg, 1969). The actual execution of the response, however, has traditionally been seen as uninteresting from a cognitive point of view. However, Abrams and Balota (1991) recently developed procedures to separate the two stages, and found that response execution processes are influenced by variables of psychological interest. Abrams and Balota required participants to complete a continuous motor movement (i.e., move a joystick) within a choice reaction time study. Within this paradigm, one can separate the time to initiate a movement (i.e., response selection time) from the speed with which that movement is completed (i.e., response execution time). Decisions associated with a greater degree of confidence not only led to a faster initiation of movement, but they also led to a faster completion of movement. Within the present context of affect and motor behavior, we note that Gray and Baruch (1987) have argued that the freezing response likely occurs within the response execution stage, rather than within the response selection stage. They reasoned that it would be inefficient for the punishment system to have inhibitory connections to the diverse set of neural systems responsible for selecting appropriate responses. It would be much more efficient for the punishment system to have one inhibitory connection to the neural area responsible for executing voluntary motor behaviors. In addition to such neurological considerations, there would also seem to be evolutionary considerations in favor of motor inhibition at the level of response execution rather than response selection. Consider a rat that detects a nearby snake. If freezing slowed response selection processes, the rat would experience difficulties choosing a course of behavioral avoidance (e.g., hiding behind a tree). This total immobility would seem to render the rat a ‘‘sitting duck,’’ so to speak. However, if freezing only slowed response execution processes, then the rat could choose a course of behavioral avoidance, but execute it in a slow and stealthful manner. As such, evolutionary considerations likely favor intact behavioral selection processes, but inhibited motor execution processes.
481
account, the current studies focused on whether motor behavior is slowed in the context of negative stimuli. To accomplish this, two paradigms required participants to respond to stimuli following positive and negative primes. We expected that under these conditions, motor facilitation and inhibition effects may be found. Study 1 required participants to first view either a positive, negative, or neutral picture, and then respond to three subsequent neutral targets. We expected that negative primes would slow reaction times to these subsequent neutral targets, whereas positive stimuli might speed such behaviors. In Study 2, we used a similar paradigm to establish that negative social cues are also capable of inhibiting subsequent motor behavior. Although such results would certainly be consistent with our hypotheses, they would not demonstrate the motor inhibition is confined to the response execution stage of processing, as hypothesized. To provide further evidence in favor of affective influences on motor execution processes, we conducted Study 3, which required participants to move a joystick in response to a directional cue following an affective prime. Following Abrams and Balota (1991), we considered the time to begin a movement a reflection of response selection processes, whereas we considered the time to complete this movement a reflection of response execution processes. As reviewed above, we predicted that inhibition and activation would be specific to the response execution stage. Study 1 In demonstrating that negatively valenced primes lead to motor inhibition, we sought to rule out a potential confound. Previous research has shown that negative stimuli tend to hold attention to a greater extent than neutral or positive stimuli do (Fox, Russo, Bowles, & Dutton, 2001; Pratto & John, 1991). To rule out anattention-related interpretation of our results, we required a series of responses subsequent to an affective prime. Theoretically, motor effects should persist beyond one initial response. If response inhibition is evolutionally designed to avoid detection by predators, as the above-cited theories suggest, it would serve little purpose if it only lasted for only one motor response. Rather, such motor inhibition effects should last long enough to avoid detection by the predator. By contrast, one would be hard-pressed to argue that slower responses on the second and third trials of the task reflect difficulties in disengaging attention from the negative prime, in that the individual has already disengaged from the prime to perform the first movement. Method
Overview of the current studies We conducted three studies to test the prediction that positive stimuli may speed motor responses, whereas negative stimuli may inhibit them. Following FanselowÕs (1994)
Participants Participants were 38 volunteers, all right-handed, who received course credit at the University of Illinois in exchange for participation.
482
B.M. Wilkowski, M.D. Robinson / Journal of Experimental Social Psychology 42 (2006) 479–490
Apparatus Two professional quality Kodak Ektapro projectors, specially outfitted with high-speed Uniblitz shutters, were connected to a dos-based computer. The shutters on both projectors were closed until the computer sent a signal for them to open. Slide onset could thus be precisely defined in terms of the signal for opening the relevant shutter. A laser pen was also interfaced with the computer, to provide a fixation point. Participants sat in a chair equipped with a small desk. A response box, with less than 1 ms random error, was attached to this desk with putty. The chair was placed 3 ft. in front of a 5 · 5 ft. projection screen. Stimuli Seventy prime slides were chosen from the International Affective Picture System (IAPS) (Lang, Bradley, & Cuthbert, 1999). These slides have been normed according valence (pleasant versus unpleasant) and arousal (high versus low) to facilitate inter-laboratory communication. Thirty positive, thirty negative, and ten neutral slides were chosen, based on the norms provided by Lang et al. (1999).1 Procedure Participants were run within individual sessions. The participants were first informed of the general nature of the experiment. They then gave informed consent and the lights were turned off. A 10-min dark adaptation period followed. Subsequently, the participants were told that they would engage in two distinct intermixed tasks. The first task consisted of viewing and processing pictorial images. To investigate the incidental effects of emotional images, we did not instruct participants to evaluate the slides. However, we did encourage them to process the images for a vaguely described memory test at the end of the study. The second task consisted of perceptual-motor performance. For this task, either one or two dots were presented. The dots were presented either to the left or right of fixation.2 Participants were asked to quickly indicate whether one or two dots were presented by pressing the appropriate button (i.e., 1 or 2) on the button box. There were three such discrimination trials before another affective prime was presented.
1
Lang et al. (1999) provide slide numbers to facilitate communication among labs. Negative stimuli consisted of slides 1050, 1111, 1120, 1220, 1300, 1301, 1930, 2053, 2520, 2800, 3130, 3230, 3250, 3350, 6230, 6260, 6300, 6510, 6570, 7361, 7380, 9008, 9290, 9300, 9320, 9410, 9415, 9421, 9561, and 9570. Positive stimuli consisted of slides 1440, 1460, 1610, 1750, 1810, 2040, 2050, 2057, 2070, 2165, 2352, 2550, 2660, 4599, 4606, 4607, 4608, 4641, 4651, 4652, 4660, 5621, 7325, 8180, 8200, 8350, 8370, 8380, 8470, and 8490. Neutral stimuli consisted of slides 6000, 7010, 7020, 7175, 7182, 7224, 7490, 7491, 7500, and 7950. 2 Images were presented to the left or right of fixation to allow the examination of emotional priming effects on lateral spatial attention (see Robinson & Compton, 2005, for related studies). This lateral presentation factor was irrelevant in the current studies.
Each trial proceeded as follows: first, the emotional slide was presented in the center of the screen for three seconds. Second, there was a 500 ms blank delay. Third, the central fixation point was presented for 200 ms. An 800 ms blank delay was then followed by 1 or 2 dots presented randomly to the left of fixation or to the right of fixation (distance from fixation = 2.5 visual angle). The dot slide was presented for 100 ms. Following the first response, there was a 1 s delay until another dot slide was presented. Although the location of the second dot trial was fixed, the number of dots was random. Following a second response, there was another 1 s delay and another dot slide, again in the same location. Following the third response, there was a 5 s delay until the presentation of the next emotional slide. There were 70 trials, one for each of the 70 slides used in the study. Each participant received a randomly generated order of emotional primes and dot stimuli. Results Reaction times and accuracy rates were each examined in a 3 (Valence: negative versus neutral versus positive) · 3 (Trial: one versus two versus three) repeatedmeasures ANOVA. Inaccurate trials (3.2% of all trials) were dropped for the purpose of analyzing reaction times. As is typical for reaction time data, the distribution was positively skewed. Thus reactions times were log-transformed to normalize the distribution (Cohen, 2001). To reduce outliers, all values 2.5 standard deviations above and below the mean (2.3% of all trials) were windsorized (Cohen, 2001). Although the analysis involved these log-transformed values, means are reported in terms of milliseconds for ease of interpretation. Analysis of accuracy rates yielded no significant effects, all ps > .10. Concerning RT, the main effect for Trial was significant, F(2,74) = 84.08, p < .0001, such that responses were slower for the first trial (M = 494 ms) than for the second (M = 424 ms) or third (M = 417 ms) trials. This reflects a task-switching cost. More importantly, however, the hypothesized main effect for Valence was significant, F(2,76) = 22.88, p < .0001, such that reaction times were slower following negative slides (M = 475 ms) than following neutral slides (M = 423 ms). Reaction times following positive slides (M = 436 ms) were actually slightly slower than the neutral baseline. To examine the time course of the Valence effect, we looked at the Valence · Trial interaction, which was non-significant, p > .60. Because Valence effects were not more pronounced on the first dot trial, such effects cannot be interpreted in terms of difficulty disengaging attention from negative primes. To further investigate the nature of the Valence main effect, we conducted two follow-up ANOVAs. One contrasted the influence of neutral and negative primes and a second contrasted the influence of neutral and positive primes. The first ANOVA indicated that negative primes slowed motor behavior relative to neutral primes, F(1,37) = 32.25, p < 0001. The second indicated that differ-
B.M. Wilkowski, M.D. Robinson / Journal of Experimental Social Psychology 42 (2006) 479–490
ence between positive and neutral primes approached, but did not reach, traditional levels of significance, F(1,37) = 3.30, p = .07. These follow-up analyses indicate that negative primes inhibited subsequent motor behavior, whereas positive primes did not facilitate motor behavior. Further analyses Much of the animal research on freezing behavior focuses on situations that are likely to induce fear (Fanselow, 1994; Fanselow & Poulos, 2005). This leaves open the question of whether freezing effects will generalize to other classes of negative stimuli besides threatening ones. In support of the generality of the negative affect/freezing relationship, we note that another variable derived from the animal conditioning literature—startle potentiation—has been tied to a wide variety of negative stimuli, including sad, disgusting, and fearful stimuli (Lang, 1995; Lang et al., 1997). Such data suggest that the present freezing effect may be primed by negative affective stimuli regardless of their specific nature. Yet, it possible that such a pattern may not be apparent here. Because fear is a high arousal type of negative affect (Lang, 1995; Russell, 1980; Tellegen, 1985), and because the negative stimuli in Study 1 systematically varied in arousal levels, we conducted a further 2 (Valence: negative versus positive) · 2 (Arousal: high versus low) · 3 (Trial: One versus two versus three) repeated measures ANOVA to address these issues. Ten slides of each level of valence and arousal level were chosen to create a fully balanced design. As in the primary analyses reported above, this analysis yielded a main effect of Valence, F(1,37) = 12.21, p = .001, and a main effect of Trial, F(1,37) = 108.94, p < .0001. However, all effects involving stimulus arousal were not significant, all ps > .13. Thus, the freezing effects found here were equally true of low and high arousal negative affective primes, a result that fits with data suggesting that numerous types of negative affect prime the neural system responsible behavioral avoidance (Lang, 1995; Lang et al., 1997). Discussion Consistent with several theories (e.g., Gray, 1987; Lang et al., 1997), Study 1 demonstrated that incidental negative primes slowed subsequent motor responses. Interestingly, positive primes also led to a slight slowing effect, although this effect was not statistically significant. We discuss possible reasons for this valence-related asymmetry (i.e., negative = slow, but positive „ fast) following Study 2. The design of Study 1 also helped to rule out an attention-related interpretation. Difficulty disengaging from negative stimuli would lead to a Valence · Trial interaction, such that the Valence effects are especially pronounced on the first dot trial. In contrast to this pattern, responses were slowed during all three dot discrimination trials. Interestingly, motor inhibition was found following all negative stimuli, whether high or low in arousal. Research
483
in the animal literature has focused somewhat exclusively on conditions that elicit fear, and freezing behavior has typically been conceptualized with respect to a fear-related context (Fanselow, 1994). Yet, paralleling the current results, startle potentiation (another variable derived from the animal literature) appears to respond to a wide variety of negative affective stimuli (Lang et al., 1997). This parallel is not surprising, given that freezing and startle potentiation are often correlated in animal conditioning studies (see Lang, 1995, p. 378). Thus, we follow Lang (1995; Lang etal., 1997) in suggesting that freezing behavior, much like startle potentiation, reflects the activation of an avoidance motivation system. This underlying motive reflects attempts to avoid negative stimuli, regardless of their specific nature (e.g., sadness-related versus fearrelated). Study 2 Study 2 sought to replicate and extend the findings of Study 1. Most theoretical discussions of freezing behavior make reference to inter-species aggression (i.e., predator– prey behavior). However, for humans living in modern, industrialized societies, inter-species aggression is a somewhat rare event, confined mostly to the experiences of recreational hunters and unlucky hikers. Although Study 1 provided some evidence that motor inhibition occurs following the presentation of a diverse set of negative images, we wished to replicate this effect within a more modern context involving frequent threats to personal safety. Specifically, we sought to demonstrate that motor inhibition occurs following the detection of intra-species (i.e., human-to-human) violence. As such, we presented participants with violent, neutral, and helpful primes. The inclusion of all three valences would again allow us to contrast the priming effects specific to negative (i.e., violent) and positive (i.e., helpful) primes. An additional focus of Study 2 was on the nature of the affective experience that is necessary to elicit motor inhibition. We know that the emotional slides used in Study 1 are generally evocative of emotional states, including effects on facial EMG, electrodermal activity, and activation of the amygdala (Lang et al., 1997). Affective words, by contrast, are less effective in creating emotional states. For example, Innes-Ker and Niedenthal (2002) recently demonstrated that having participants unscramble affect-laden sentences increased the accessibility of mood-consistent concepts, as measured by a lexical decision task, but did not influence mood states. Thus, effects primed by valenced words are presumably due to affective evaluations rather than subjective feeling states. In this connection, we sought to determine whether negative words would be capable of inducing freezing behavior in the present paradigm. If so, the results cannot be attributed to changes in subjective feeling from trial to trial. Rather, such effects would be attributable to incidental evaluations.
484
B.M. Wilkowski, M.D. Robinson / Journal of Experimental Social Psychology 42 (2006) 479–490
Procedure All sessions included up to five participants. Participants came to the lab, gave informed consent, and were told of the general nature of the study by the experimenter. The experimenter then started all participants on the computer program, which included more detailed instructions. As in the previous study, participants were told that there were two interwoven components to the task. The first involved memorizing words and the second involved a perceptual/ motor task. Participants first saw a prime word at the center of the computer screen, which they were to memorize. These prime words consisted of violent, neutral, and helpful words.3 The prime stayed on screen for 1 s. Following a 200 ms pause, the letter ‘‘p’’ or ‘‘q’’ appeared two lines above or below the prime. Participants were asked to press the appropriate button on the keyboard (i.e., p or q) as quickly and as accurately as possible. The letter stayed on screen until a response was registered. Incorrect trials were followed by a 2 s error message. Correct trials were followed by a 150 ms pause. Parallel to Study 1, there were two additional trials involving letter identification. This letter appeared in the same location, but the identity of the letter was chosen randomly. Two hundred milliseconds following the third trial, participants were prompted to type in the prime word. There was then a 2 s pause until the next prime. Each word was shown twice, resulting in 90 trials. Each participant received a randomly generated sequence of primes, q/p targets, and target locations.
third (M = 98.69) trials. A marginal Trial · Valence interaction was also found, F(4,160) = 2.12, p = .08. The pattern of means for the latter interaction was quite irregular, and appeared to be due to a differential effect of violent versus helpful primes on the second q/p trial. Given the marginal nature of this interaction, we discounted it as anomalous. All inaccurate trials (1.67% of trials) were dropped for purposes of RT analysis. To correct for a skewed distribution, all RTs were log-transformed, and outliers 2.5 SDs above or below the mean were windsorized (1.64% of trials). Although analyses were conducted in terms of these log-transformed values, means are reported in the original units for ease of interpretation. A main effect for Trial was found, F(2,80) = 446.52, p < .0001, such that RTs were slower on the first trial (M = 862 ms) than on the second (M = 499 ms) or third (M = 502 ms) trials. This reflects the cost of switching from the word memorization task to the letter identification task. More importantly, the hypothesized main effect for Valence was also significant, F(2,80) = 7.42, p < .02, such that responses were slower following violent primes (M = 633 ms) than following neutral primes (M = 618). RTs following helpful primes (M = 611 ms) were slightly faster than the neutral condition baseline. We examined the time course of the Valence effect by looking at the Valence · Trial interaction, which was non-significant, p > .53. Thus, results cannot be interpreted in terms of a difficulty disengaging attention from violent words, which would mainly affect the first trial following the word memorization priming task. To examine the nature of our Valence main effect, we conducted two follow-up ANOVAs. The first contrasted violent and neutral primes, and found there was a reliable slow-down following violent primes, F(1,40) = 7.57, p < .01. The second ANOVA contrasted helpful and neutral primes, and found there was no significant difference between the two prime types, p = .38. Thus, violent primes slowed subsequent motor behavior, but helpful primes did not speed subsequent motor behavior.
Results
Discussion
Reaction times and accuracy rates were analyzed in a 3 (Valence: violent versus neutral versus helpful) · 3 (Trial: one versus two versus three) repeated measures ANOVA. Analysis of accuracy rates indicated two marginal effects, a main effect for Trial, F(2,80) = 2.51, p = .08, such that accuracy rates were slightly lower on the first trial (M = 98.07%) than on the second (M = 98.26%) or the
Study 2 extended results related to motor inhibition by demonstrating that motor inhibition occurs following the detection of violent word cues (Study 2) as well as pictures of a more general negative type (Study 1). An additional focus of Study 2 was to demonstrate inhibition effects due to negative words rather than negative images. As affective words do not produce strong feeling states (Innes-Ker & Niedenthal, 2002), we suggest that the mere detection of a negative stimulus is sufficient to cause motor inhibition effects. Helpful stimuli, like the more general category of positive stimuli in Study 1, did not facilitate subsequent motor behaviors. This finding conceptually replicates Study 1, as well as the findings of Derryberry (1991), and clearly suggests that motor inhibition following negative primes
Methods Participants Participants were 41 undergraduate psychology students at North Dakota State University who participated in exchange for extra credit. Apparatus All sessions were completed on one of five Windows based computers using E-Prime software.
3
Violent words were as follows: attack, choke, decapitate, kill, massacre, murder, mutilate, punch, rape, shoot, shove, slaughter, stab, suffocate, and torture. Neutral words were as follows: add, arrange, compose, connect, copy, crawl, fasten, indent, make, sort, stack, tape, utilize, walk, and weave. Helpful words were as follows: befriend, care, comfort, give, help, hug, love, nurture, praise, relieve, respect, smile, sooth, sympathize, and trust.
B.M. Wilkowski, M.D. Robinson / Journal of Experimental Social Psychology 42 (2006) 479–490
is a more robust phenomenon than is motor facilitation following positive primes. One potential conclusion from this is that motor facilitation is non-existent. However, considering results from the animal neuroscience literature (see Depue & Collins, 1999), we believe this conclusion to be premature. Depue and Collins (1999) suggest that motor facilitation effects due to positive affect tend to occur only within certain situations, specifically those associated with a moderate, rather than high, desire to perform the behavior in question. We can reasonably assume our participants were highly motivated to perform the movement indicated, in that it was a necessary part of finishing the task. Under other conditions, in which participants are less motivated to perform the behavior, positive affect may in fact facilitate motor behavior (Depue & Collins, 1999). Future studies will be necessary to examine potential motor facilitation effects following positive primes. This said, we now turn the focus of the present research to exploring the nature of motor inhibition in more detail. Studies 1 and 2 had a few limitations that are worth mentioning. First, they did not measure motor behavior continuously, but rather used successive trials to examine the time course of motor behavior. This is a relatively indirect way to measure the speed of motor behavior, and it is somewhat silent on whether the effects occur prior to the selection of a response or subsequent to the initiation of this response. That is, keys on a computer keyboard respond in an all-or-none manner to pressure and therefore cannot separate the initiation of pressure from the completion of pressure; both are defined by the same event, namely sufficient key pressure. To better examine the nature of motor behavior, we used a joystick in Study 3, thus rendering it possible to separate the initial onset of movements from the subsequent response execution processes. This would allow us to explore the hypothesis that motor inhibition effects would be confined to the response execution stage (see Gray & Baruch, 1987). An additional limitation of Studies 1 and 2 was that responses always involved button presses. Button presses can, at least under certain conditions with certain stimuli, be implicitly represented as an approach-related behavior (Wentura, Rothermund, & Bak, 2000; Study 3). This renders it possible that these effects reported above may have involved directional motor effects (i.e., negative = slower approach) rather than non-directional motor effects (i.e., negative = slower behavior). Study 3 sought to change the response paradigm to separate directional and non-directional effects. We did this by requiring both forward and backward directions of movement in Study 3. Study 3 In this study, participants were first presented with an affective prime, and then asked to respond to a directional cue (forward or backward) with the appropriate joystick movement. Directional cues were randomly assigned to
485
word trials and thus bore no relationship to the nature of the particular affective primes presented. As such, the affective primes did not signal the particular movement to be performed, nor were they informative in any manner. Given these task conditions, effects on motor behavior would be purely incidental in nature. To differentiate between response selection and response execution processes within such a paradigm, we borrowed procedures from Abrams and Balota (1991), who used a continuous measurement of choice reaction time. With a continuous measurement, we could separate the time required for the initiation of movement (i.e., response selection) from speed to complete the motor behavior itself (i.e., response execution). We reasoned, following the material presented in the introduction, that inhibition effects due to a negative prime might be particular to the response execution stage rather than the response selection stage. Method Participants Participants were 26 students at the University of Illinois who participated in exchange for course credit. Apparatus The computer, projectors, and seating arrangement were identical to Study 1. The first projector displayed valenced words and the second projector displayed the letters F (Forward) or B (Backward), depending on trial. Study 3 used a joystick rather than a response box. The joystick was specifically constructed for this project. An 8-in. metal tube was encased within a wooden box. A slit within the box allowed for 6 in. of movement in either the forward or the backward direction. Steel springs held the joystick in place, whereas a string gauge converted position data into a voltage signal that could be read by the computer. Four thousand unique positions were registered and position was sampled every 33 ms. Stimuli Eighty valenced words were taken from Osgood, May, and Miron, 1975, Appendix F. These stimuli were rated in 23 separate communities according to several factors, including evaluation. Studies 1 and 2 suggested that positive stimuli were equivalent to neutral stimuli with respect to their effects on motor behavior. We thus decided to drop the neutral condition from Study 3. In retrospect, we wish that we had included a neutral condition, as we did in Studies 1 and 2, to better document the equivalence of positive and neutral primes within a different behavioral paradigm. We therefore suggest that the issue of freezing (due to negative stimuli) versus facilitation (due to positive stimuli) is necessarily a somewhat open one with respect to the design of Study 3 (but not Studies 1 and 2). Study 3 used 40 positive and 40 negative words that were relatively equal in levels of emotional arousal (or
486
B.M. Wilkowski, M.D. Robinson / Journal of Experimental Social Psychology 42 (2006) 479–490
Recorded Points Movement Onset Movement Termination 2000
Joystick Position
1500 1000 500 0 -500 0
500
1000
1500
2000
2500
-1000 -1500 -2000 Time Fig. 1. Computer algorithms and motor performance: a sample trial from Study 3.
potency, in the Osgood et al., 1975, scheme).4 To demonstrate that OsgoodÕs 40 year-old norms are not outdated, we also had six naı¨ve graduate students evaluate these words on a continuous bipolar evaluation scale. These evaluations were highly correlated with OsgoodÕs original norms, r = .90, demonstrating that these Osgood et al.Õs (1975) norms are still quite valid. Procedure As in Study 1, participants were tested within individual sessions. Participants were informed of the general nature of the procedures and then dark-adapted prior to performance in the experimental task. Participants were told that they would perform two intermixed tasks. First, they would see words that they should attempt to memorize. Second, they would perform an indicated movement, either in a forward or backward direction, using a joystick. This movement should be performed as quickly and accurately as possible. As in previous studies, we did not make any mention of the manipulated valence of the prime stimuli. However, as in prior studies, we did encourage focused processing of the prime. Each trial consisted of the following sequence: The valenced word appeared for 2 s. During this time the participants began rehearsing the word. After 2 s, a 4 Negative words were as follows: cancer, earthquake, disease, filth, thief, sickness, prison, anger, snake, failure, death, defeat, sadness, battle, fighting, noise, funeral, shame, baldness, beggar, deaf, crow, fat, insect, urine, lion, widow, fire, stone, weight, rock, wall, passion, power, salt, eternity, elephant, steel, floor, law, and pyramid. Positive words were as follows: mother, happiness, love, spring, flower, music, kissing, health, friend, beauty, bride, pleasure, girl, house, library, bed, baby, youth, courage, wedding, sleep, hero, champion, tree, moon, pillow, automobile, laughter, dancing, sex, land, forest, rug, map, table, electricity, mountains, jazz, and joke.
movement cue was presented below the word. The movement cue was either the letter F for forward of B for backward. Simultaneous with the appearance of the movement cue, we began sampling joystick position. Position was sampled for 2 s, once every 33 ms. After the 2 s recording period, participants said the word out loud and then there was a 5 s delay until the appearance of the next word. There were 80 trials, one for each of the word stimuli. Each participant received a randomly generated sequence of prime stimuli and directional cues. Results Data reduction Participants had been instructed to hold the joystick throughout the course of the study. Because of alterations in body posture and fluctuations in muscular control, joystick position varied on a more or less continuous basis. We sought to create computer algorithms to determine when participants began and ended movement in the indicated diretion. These algorithms are explained in more detail in Appendix A. Fig. 1 is a graphic display of one trial in the study, and displays the time and position of the onset and termination of the movement, as extracted by the computer program. The x-axis displays time since the appearance of the directional cue. The y-axis displays the position of the joystick. In this trial, the participant was instructed to move the joystick backwards. As the graph shows, both the onset and termination of the movement in the indicated direction are fairly obvious to the naked eye. However, given the large number of trials involved and the desire for an objective, invariant procedure, we created computer algorithms to recover the parameters of key interest (see Appendix A).
B.M. Wilkowski, M.D. Robinson / Journal of Experimental Social Psychology 42 (2006) 479–490
Error rates Two participants were found to have unacceptably high error rates (>40%), and as such a performance indicates an only slightly better than chance performance, their data was dropped. Among the remaining participants, 6.76% of trials were not capable of analysis by the computer algorithms. A visual inspection of a subset of these trials revealed that the vast majority of them were highly irregular and uninterpretable to the human eye as well. With the removal of these trials, 99.19% of the movements were in the correct direction. Analysis of variance revealed no effects of Valence, Direction, or their interaction on the accuracy of directional movements, all ps > .10. Movement onset All inaccurate trials were dropped from the analysis of movement onset. Once again, a positively skewed distribution was found. Onset times were log-transformed to normalize the distribution (Cohen, 2001). We analyzed movement onset times in a 2 (Valence: positive versus negative) · 2 (Direction: forward versus backward) repeated measures ANOVA. No significant effects were found, all ps > .10. We discuss such null results more completely in the discussion of this study. Velocity Velocity scores were calculated by dividing distance of movement (i.e., position at movement offset position at movement onset) by time of movement (i.e., time at movement offset time at movement onset). The distribution of velocities was also skewed and we therefore log-transformed them for analysis (Cohen, 2001). However, means are reported in terms of the original untransformed velocity units for ease of interpretation. Velocities were analyzed in a 2 (Valence: positive versus negative) · 2 (Direction: forward versus backward) repeated measures ANOVA. The analysis revealed a significant main effect for Valence, F(1,23) = 4.93, p < .04. Movements following positive primes were faster (M = 43.35 cm/s) than movements following negative primes (M = 42.48 cm/s). In other words, following positive words, participants moved a greater distance per second. No other effects in the analysis were significant, all Fs < 1. Further analyses Many of the affective primes used in the present study were typically not extreme in nature. When this fact is considered in light of our reliance on 40-year old word norms (i.e., from Osgood et al., 1975), it may be somewhat desirable to replicate the valence effect using a subset of affective primes that are more clearly valenced in nature. We determined that the most extreme 23 of each valence were clearly valenced in nature. We subsequently re-ran the analysis using only this subset of the 23 most positive and 23 most negative primes. Similar to the analyses reported above, we found no effects on movement accuracy or movement onset times, ps > .10. However, an analysis of movement veloci-
487
ties replicated the primary analysis reported above, in that movement velocities were slower following negative primes (M = 42.72 cm/s) than following positive primes (M = 43.84 cm/s), F(1,23) = 4.66, p < .05. No other effects within the analysis of velocities were significant, Fs < 1. Therefore, the effects of more obviously valenced primes replicated the full analysis reported above in revealing that motor behavior is slowed in the context of negative affective primes. Discussion The speed of motor execution is typically ignored within cognitive (and certainly affective) models of processing. One very practical reason for this oversight is that traditional measures of reaction time, which involve discrete actions, cannot disentangle response selection and response execution processes (Abrams & Balota, 1991). Study 3 was the first study, to our knowledge, to look at response execution within an incidental evaluation task. Its findings clearly support the prediction of motivational theories of emotion (e.g., Gray, 1987; Lang et al., 1997). That is, voluntary behavior was inhibited in the context of negative affective stimuli. Consistent with theoretical speculation and neuropsychological results (Gray & Baruch, 1987), these effects were confined to the stage of response execution (as revealed by the velocity measure), and did not appear in the stage of response selection (as revealed by results involving movement onset times). As mentioned above, such results are consistent with both neurological and evolutionary concerns. From a neurological point of view, it is much easier for the punishment system to have one inhibitory connection to the neural area responsible for response execution, rather than have numerous inhibitory connections to the many areas involved in response selection (Gray & Baruch, 1987). Likewise from an evolutionary viewpoint, an inhibition of response selection processes may impede movements to avoid detection by predators (e.g., hiding behind a tree). However, an inhibition of response execution processes would not impede such avoidance decisions, but would rather insure that avoidance behaviors, once initiated, are performed in a stealthful and thus less detectable manner. General discussion Summary of findings Study 1 provided preliminary evidence that incidental exposure to valenced stimuli can influence the speed of subsequent motor behavior. Reaction times following negative stimuli were slower across three consecutive trials in comparison to reaction times following neutral stimuli. In contrast, reaction times following positive primes were not significantly different than baseline. As discussed above, this suggests that response facilitation following positive primes is a less robust phenomenon, but it does not rule out the possibility that human participants will exhibit such
488
B.M. Wilkowski, M.D. Robinson / Journal of Experimental Social Psychology 42 (2006) 479–490
effects within other paradigms in which the participant is less motivated to respond to the target (see Depue & Collins, 1999). Study 2 extended these results by showing that violent primes also led to a motor inhibition effect, whereas helpful stimuli did not facilitate motor responses. Although the findings from Studies 1 and 2 were novel, robust, and clearly supportive of our hypothesis, we felt the need to conduct a third study to examine two issues. First, the use of discrete RTs was insufficient for separating response selection processes from response execution processes. Second, the use of only one movement direction (specifically, a button press) renders it possible that the effects were direction-specific, as there is at least a small amount of evidence suggesting that button presses can sometimes be interpreted in terms of approach-related behavior (Wentura et al., 2000; Study 3). Study 3 sought to eliminate these ambiguities. By measuring choice RT in a continuous fashion, we could separate response selection processes from response execution processes. We also varied the direction of the movements. As hypothesized, movement velocities were slower in the context of negative stimuli. Thus, evidence from three studies supports the notion that motor behavior is slowed within the context of negative affective primes. Implications for motivational theories of affect Our results further the case of motivational theories of affect, which postulate that a reward and a punishment motivation system underlie affective reactions to stimuli (e.g., Cacioppo et al., 1999; Davidson, 1992; Gray, 1987). Many such theories argue that these systems have their effect by priming particular motor responses that should be adaptive within the context of reward pursuit or punishment avoidance. Some prior evidence for such motor priming effects comes from studies showing that, under certain conditions, positive stimuli may facilitate approach-related behaviors, whereas negative stimuli may facilitate avoidance-related behaviors (Neumann et al., 2003). The current studies add to this picture by showing that, under different conditions, non-directional motor priming effects also occur, such that voluntary behavior is inhibited by negative primes. We suggest that FanselowÕs (1994) analysis of defensive behaviors may be useful in distinguishing the directional motor priming effects demonstrated in previous research (e.g., Chen & Bargh, 1999) from the non-directional priming effects found in the current studies. First, Fanselow argues that when one responds directly to a negative stimulus, flight or fight behaviors are often primed. Consistent with this, prior paradigms relating affect to approach and avoidance behaviors (e.g., Chen & Bargh, 1999; Markman & Brendl, 2005) required participants to respond directly to a valenced stimulus. Second, Fanselow argues that when one is in the vicinity of a negative stimulus, but not directly responding to it, freezing behavior results. Consistent with this account, the current studies made
movement contingent upon neutral stimuli occurring within the context of a negative stimulus, and found motor inhibition effects. Together, studies suggest that although negative stimuli may somewhat inevitably activate the aversive motivational system, the exact motor behavior triggered by this system depends upon the nature of the situation. That is, the aversive motivational system takes into account the nature of the situation before priming any specific motor behavior. Thus, this system is actually quite flexible in giving rise to freezing, fleeing, or fighting behavior depending on the imminence of a negative stimulus (Fanselow, 1994; Lang et al., 1997). Together, then, the present and prior studies (e.g., Chen & Bargh, 1999) help to provide a more complete account of the motor responses linked to affect, by grounding theories based mostly on animal neuroscience (e.g., Gray, 1987) in behavioral research with human participants. The current paradigm may therefore prove to be useful in linking animal research using the freezing response (Fanselow & Poulos, 2005) to theoretical conceptualizations of the human affective system. The nature of the motivational systems Study 1 found that motor inhibition occurred with respect to both low arousal (i.e., sadness-inducing) and high arousal (i.e., fear- and disgust-inducing) negative stimuli. This pattern of data suggests that motor inhibition may be a consequence of the aversive motivational system rather than any specific discrete negative emotion. However, a number of recent theories (Carver, 2004; Harmon-Jones, 2003; Higgins, 1997) have suggested that at least some forms of negative affect cannot be mapped onto the aversive motivational system. These theorists contend that, at least in some instances, anger and sadness can result from difficulties involved in obtaining rewards. In these cases, the appetitive (rather than the aversive) system may be the driving force behind anger and sadness (Carver, 2004; Harmon-Jones, 2003; Higgins, 1997). Such theories are directly opposed to more traditional views of the motivational systems underlying affect, which suggest that all forms of negative affect reflect the broad operation of an aversive motivational system (e.g, Gray, 1987; Lang et al., 1997). These two different frameworks, although grounded in motoric conceptions of affect, have not presented relevant motoric evidence. The present paradigms should be capable of providing evidence for the motoric basis of emotional experiences like anger or sadness. Traditional theories (e.g., Gray, 1987; Lang et al., 1997) suggest that the absence of an expected reward should result in motor inhibition. More recent theories (Carver, 2004; Harmon-Jones, 2003; Higgins, 1997), by contrast, suggest such an absence should not trigger motor inhibition. By combining the current paradigms with frustration-inducing versions of conditioning procedures (see Dollard, Doob, Miller, Mowrer, & Sears, 1939), these different predictions can be tested. In
B.M. Wilkowski, M.D. Robinson / Journal of Experimental Social Psychology 42 (2006) 479–490
sum, the current methodology provides a means of clarifying how the motivational systems underlying affect are conceptualized. Conclusions The present studies examined whether incidental exposure to negative stimuli is capable of slowing the speed of voluntary motor behavior. The results indicate that the answer is yes, and that current motivational theories (e.g., Gray, 1987; Lang et al., 1997) have relevance to the incidental evaluation of affective stimuli. The findings therefore link theories of automatic evaluation to neurocognitive and motivational theories of affect. Acknowledgment The authors acknowledge grant support from NIMH (MH 068241).
489
less dramatic movement onsets, as might be missed by algorithm A. As such, algorithm B was sufficient to determine movement onset. At this point, the program once again scrolls forward, and applies algorithm C until such time as it returns a value less than 20. This algorithm was designed to detect a leveling out trend (i.e., the cessation of movement). In most cases, algorithm C was sufficient. However, algorithm C was only sensitive to leveling-out trends of at least four recordings. In the case of such rapid returns, C would yield and error message and algorithm D would be performed. Algorithm D was more sensitive to briefer leveling out trends occurring over the course of three consecutive samples. In combination, the results of algorithms C and D were sufficient to determine movement offset. Movement onset and offset values were used to determine the velocity of movement (i.e., [position of offset position of onset]/ [time of offset time of onset]). References
Appendix A. Data reduction algorithms used in study 3 The following account is meant to give a brief overview of the functioning of the program so as to make the process accessible to the reader. It is not sufficient to completely describe the functions of the program in every detail. Four algorithms were used to extract movement onset and movement termination: A ¼ j P n P n1 P n2 j; where Pn is the current joystick position and Pn1 is the joystick position at last reading. B ¼ j ðP n P n4 Þ þ ðP n1 P n5 Þ þ ðP n2 P n6 Þ þ ðP n3 P n7 Þ j; C ¼ kðP n þ P nþ1 þ P nþ2 þ P nþ3 Þ=4 j j P n k þ kðP nþ1 þ P nþ2 þ P nþ3 þ P nþ4 Þ=4 j j P n k; D ¼ j P n þ P nþ1 þ P nþ2 j j P nþ3 þ P nþ4 þ P nþ5 j . The program was structured such that it scrolled through the data of an individual trial in chronological order, applying the algorithms in sequence until a specific value was reached. Algorithm A is applied until the returning value (related to position change) exceeded 1000. This algorithm was designed to indicate a large difference in joystick positions over the span of three data points. It was designed to be insensitive to small fluctuations in movement, but to identify position data that quite unambiguously indicates a voluntary movement. The algorithm sometimes misses less extreme voluntary movements. The second algorithm is used to correct for the relative insensitivity of algorithm A. The algorithm works backwards from the position returned by A by applying algorithm B until the returning value is below 250. This algorithm is designed to identify a string of several joystick positions with little variation. This criterion is sufficiently sensitive to relatively
Abrams, R. A., & Balota, D. A. (1991). Mental chronometry: Beyond reaction time. Psychological Science, 2, 153–157. Bargh, J. A. (1989). Conditional automaticity: Varieties of automatic influence in social perception and cognition. In J. S. Uleman & J. A. Bargh (Eds.), Unintended thought (pp. 3–51). New York: Guilford Press. Cacioppo, J. T., Gardner, W. L., & Bernston, G. G. (1999). The affect system has parallel and integrative processing components: Form follows function. Journal of Personality and Social Psychology, 76, 839–855. Carver, C. S. (2004). Self-regulation of action and affect. In R. F. Baumeiser & K. D. Vohs (Eds.), Handbook of self-regulation (pp. 13–39). New York: The Guilford Press. Carver, C. S., Sutton, S. K., & Scheier, M. F. (2000). Action, emotion, and personality: Emerging conceptual integration. Personality and Social Psychology Bulletin, 26, 741–751. Chen, M., & Bargh, J. A. (1999). Consequences of automatic evaluation: Immediate behavioral predispositions to approach or avoid the stimulus. Personality and Social Psychology Bulletin, 25, 215–224. Cohen, B. H. (2001). Explaining psychological statistics. New York: Wiley. Crawford, F. A., & Masterson, M. (1982). The defense motivation system: A theory of avoidant behavior. The Behavioral and Brain Sciences, 5, 661–696. Davidson, R. J. (1992). Emotion and affective style: Hemispheric substrates. Psychological Science, 3, 39–43. Depue, R. A., & Collins, P. F. (1999). Neurobiology of the structure of personality: Dopamine, facilitation of incentive motivation, and extraversion. Behavioural and Brain Sciences, 22, 491–569. Depue, R. A., & Lenzenweger, M. F. (2001). A neurobehavioral dimensional model. In W. J. Livesley (Ed.), Handbook of personality disorders: Theory, research, and treatment (pp. 136–176). New York: Guilford Press. Derryberry, D. (1991). The immediate effects of positive and negative feedback signals. Journal of Personality and Social Psychology, 61, 267–278. Dollard, J., Doob, L. W., Miller, N. E., Mowrer, O. H., & Sears, R. R. (1939). Frustration and aggression: Fundamental question. New York: Oxford University Press. Ekman, P., & Davidson, R. J. (1994). The nature of emotion: Fundamental questions. Oxford, UK: Oxford University Press. Fanselow, M. S. (1994). Neural organization of the defensive behavior system responsible for fear. Psychonomic Bulletin and Review, 1, 429–438.
490
B.M. Wilkowski, M.D. Robinson / Journal of Experimental Social Psychology 42 (2006) 479–490
Fanselow, M. S., & Poulos, A. M. (2005). The neuroscience of mammalian associative learning. Annual Review of Psychology, 56, 207–234. Fazio, R. H., Sanbonmatsu, D. M., Powell, M. C., & Kardes, F. R. (1986). On the automatic activation of attitudes. Journal of Personality and Social Psychology, 50, 229–238. Fowles, D. C. (1980). The three arousal model: Implications of GrayÕs two-factor learning theory for heart rate, electrodermal activity, and psychopathy. Psychophysiology, 17, 87–104. Fox, E., Russo, R., Bowles, R., & Dutton, K. (2001). Do threatening stimuli draw or hold visual attention in subclinical anxiety? Journal of Experimental Psychology: General, 130, 681–700. Gray, J. A. (1987). The neuropsychology of emotion and personality. In S. M. Stahl, S. D. Iversen, & E. C. Goodman (Eds.), Cognitive neurochemistry (pp. 171–190). Oxford, UK: Oxford University Press. Gray, J. A., & Baruch, I. (1987). DonÕt leave the ‘‘psych’’ out of neuropsychology. Behavioral and Brain Sciences, 10, 215. Grice, G. R., Nullmeyer, R., & Spiker, V. A. (1982). Human reaction time: Towards a general theory. Journal of Experimental Psychology: General, 111, 136–153. Harmon-Jones, E. (2003). Clarifying the emotive functions of asymmetrical frontal cortical activity. Psychophysiology, 40, 838–848. Higgins, E. T. (1997). Beyond pleasure and pain. American Psychologist, 52, 1280–1300. Innes-Ker, A., & Niedenthal, P. M. (2002). Emotion concepts and emotional states in social judgment and categorization. Journal of Personality and Social Psychology, 83, 804–816. Klauer, K. C., & Musch, J. (2003). Affective priming: Findings and theories. In J. Musch & K. C. Klauer (Eds.), The psychology of evaluation: Affective processes in cognition and emotion (pp. 7–50). Mahwah, NJ: Lawrence Erlbaum. Lang, P. J. (1995). The emotion probe: Studies of motivation and attention. American Psychologist, 50, 372–385. Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1997). Motivated attention: Affect, activation, and action. In P. J. Lang, R. F. Simons, & M. T. Balaban (Eds.), Attention and orienting: Sensory and motivational processes (pp. 97–135). Mahwah, NJ: Lawrence Erlbaum. Lang, P. J., Bradley, M. M., Cuthbert, B. N. (1999). International affective picture system (IAPS): Technical manual and affective ratings. The Center for Research in Psychophysiology, University of Florida, Gainesville, FL. Markman, A. B., & Brendl, C. M. (2005). Constraining theories of embodied cognition. Psychological Science, 16, 6–10.
McClelland, J. L. (1979). On the time relations of mental processes: An examination of systems of processes in cascade. Psychological Review, 86, 287–330. Musch, J., & Klauer, K. C. (2001). Locational uncertainty moderates affective congruency effects in the evaluative decision task. Cognition and Emotion, 15, 167–188. Neumann, R., Fo¨rster, J., & Strack, F. (2003). Motor compatibility: The bidirectional link between behavior and evaluation. In J. Musch & K. C. Klauer (Eds.), The psychology of evaluation (pp. 371–392). Mahwah, NJ: Lawrence Erlbaum. Osgood, C. E., May, W. H., & Miron, M. S. (1975). Cross-cultural universals of affective meaning. Urbana, IL: University of Illinois Press. Pavlov, I. P. (1927). Conditioned reflexes. New York, NY: Dover. Pessoa, L., McKenna, M., Gutierrez, E., & Ungerleider, L. G. (2002). Neural processing of emotional faces requires attention. Proceedings of the National Academy of Sciences USA, 99, 11458–11463. Pratto, F., & John, O. P. (1991). Automatic vigilance: The attentiongrabbing power of negative social information. Journal of Personality and Social Psychology, 61, 380–391. Robinson, M. D., Compton, R. J. (2005). The automaticity of affective reactions: Stimulus valence, arousal, and lateral spatial attention, submitted. Russell, J. (1980). A circumplex model of affect. Journal of Personality and Social Psychology, 39, 1161–1178. Sternberg, S. (1969). The discovery of processing stages: Extensions of DonderÕs method. Acta Psychologica, 30, 276–315. Tellegen, A. (1985). Structures of mood and personality and their relevance to assessing anxiety, with an emphasis on self-report. In A. H. Tuma & J. D. Maser (Eds.), Anxiety and the anxiety disorders. New Jersey: Erlbaum. Watson, D. (2000). Mood and temperament. New York, NY: Guilford Press. Watson, D., Wiese, D., Vaidya, J., & Tellegen, A. (1999). The two general activation systems of affect: Structural findings, evolutionary considerations, and psychobiological evidence. Journal of Personality and Social Psychology, 76, 820–838. Wentura, D., Rothermund, K., & Bak, P. (2000). Automatic vigilance: The attention- grabbing power of approach- and avoidance-related social information. Journal of Personality and Social Psychology, 78, 1024–1037. Wentura, D., & Rothermund, K. (2003). The ‘‘meddling-in’’ of affective information: A general model of automatic evaluation. In J. Musch & K. C. Klauer (Eds.), The psychology of evaluation: Affective processes in cognition and emotion (pp. 51–86). Mahwah, NJ: Lawrence Erlbaum.
