Abstract - University of North Carolina Wilmington



Negotiating affective steering currents; Cognitive Bias

This report examines and attempts to integrate concepts and studies relevant to cognitive bias from a range of scientific disciplines and subfields. Taken collectively It is argued here that cognitive bias is the expression of underlying motivational/affective states that automatically and unconsciously produce “steering currents” for information processing that are negotiated or tempered by controlled processing. These affective steering currents (cognitive bias) may dominate cognitive processing when the affective state is particularly strong, when affectively relevant stimuli are particularly intense, or when controlled processing is diminished. This affective steering current theory elaborates upon the wealth of evidence suggesting that different motivational states may influence cognition at different points of processing (from attention, encoding, interpretation, judgment, schema, stereotype, values, and behavior), and consequently may automatically influence effortful processing from the level of attention and encoding (Lundh, Wikström, and Westerlund, 2001), through higher order goals, schemas and beliefs (Dovidio, Kawakami, & Gaertner, 2002). Moreover, explicit, elaborative processing is likely to involve secondary implicit, automatic processes that can bias their direction and degree. This theoretical view is consistent with others who propose that emotion and cognition are interactive processes and that affect may influence both early (perception and attention) and late cognitive operations (Zajonc,1980; Seminowicz and Davis; Storbeck and Clore, 2007; MacDonald, 2008; Duncan and Barrett, 2007; Bower, 1981; Bruner, 1957; Brown, Scott, Bench, & Dolan, 1994; Dalgleish & Power, 2000; Forgas, 1995; J. R. Gray, 1999; Heller, 1990; Heller & Nitschke, 1998; Humphreys & Revelle, 1984; Isen, 1993; Koelega, 1992; Oaksford, Morris, Grainger, & Williams, 1996; Revelle, 1993). The “steering current” theory also develops an extensive set of empirically testable predictions.

Cognitive Bias; implicit vs explicit processes

Cognitive bias is defined as an automatic vigilance to motivationally relevant stimuli that deter the effortful processing of other stimuli ( Gronau, Cohen, & Ben-Shakhar, 2003; Hartikainen, Ogawa, & Knight, 2000; MacDonald, 2008). Such cognitive bias is argued to result from increased stimulus salience due to emotional or motivational influences on cognitive processes (Zajonc,1980; White, 1996; Okon-Singer, Hadas, Tzelgov, Joseph, Henik, Avishai, 2007). It is obvious that bias in information processing may occur as an intentional act or strategy, but emphasis here is on the biases that appear to occur automatically, without consciousness (MacLeod, Rutherford, Cambell, Ebsworthy and Holker, 2002; Tzelgov,Porat & Henik 1997) and without intention (Fox, 1996; MacLeod & Rutherford (1992) C. MacLeod and E. Rutherford, Anxiety and the selective processing of emotional information: Mediating roles of awareness, trait and state variables, and personal relevance of stimulus materials, Behaviour Research and Therapy 30 (1992), pp. 479–491. Abstract | View Record in Scopus | Cited By in Scopus (152)MacLeod & Rutherford, 1992; Hasher & Zacks, 1979; Schneider, Dumais, & Shiffrin, 1984).

Several theories of information processing describe cognitive bias as an implicit process. Implicit and explicit mechanisms may be contrasted on a number of dimensions (e.g.,Geary, 2005; Lieberman, 2007; Satpute & Lieberman, 2006; Stanovich, 1999). Implicit processing is automatic, effortless, relatively fast, and independent from consciousness and intention (Reber, 1993; Remillard and Clark, 2001; Hasher & Zacks, 1979; Schneider, Dumais, & Shiffrin, 1984; Sherman, Gawronski, Gonsalkorale, Hugenberg, Allen and Groom, 2008; but see Okon-Singer, Hadas, Tzelgov, Joseph, Henik, Avishai, 2007). Implicit processes related to motivational state that underline cognitive biases appear to exist and influence a broad range of functions from attention, memory and perceptual interpretations (e.g., stereotypes), to priming effects (Bargh& Chartrand, 1999), and behaviors that have become automatic with repetition. Integrative and data-driven accounts of implicit processing argue that this automatic process enhances stimulus properties of an event without the strategic elaboration or activation of associative networks that occurs with explicit, controlled cognitive processes ((Blaxton, 1989; Jacoby, 1983; Roediger & Blaxton, 1987b; Williams, Watts, MacLeod, and Mathews 1988; Graf and Mandler, 1984).

Explicit cognitive processing involves high level functions over which we may theoretically exert conscious, intentional control, conscious retrieval, elaboration (Mischel, 1984; Mischel, Cantor, & Feldman, 1996), and goal commitment (Gollwitzer, 1990; Mischel et al., 1996; Muraven & Baumeister, 2000). These high order processes have often been referred to differently by different investigators. Executive function (Eysenck, Derakshan, Santos and Calvo, 2007; Friedman and Miyake, 2004), “central executive” (Baddeley, 2001; Duchek et al. 1998; e.g.,Spieler et al. 1996; Zacks et al. 1996; West and Bell 1997; May et al. 1999; Gazzaley & D’Esposito, 2008), self-control (Fishbach and Shah, Dhar & Wertenbroch, 2000; Freitas, Liberman, & Higgins, 2002; Gollwitzer, 1999; Kivetz & Simonson, 2002; Kuhl, 1986; Loewenstein, 1996; Metcalfe & Mischel, 1999; Muraven & Baumeister, 2000; Trope & Fishbach, 2000), cognitive control (Fernandez-Duque and Knight, 2008;Gray, 2001; Braver & Cohen, 2000; Posner & Snyder, 1975; Logan, 1985; Posner & DiGirolamo, 1998; Posner & Fernandez-Duque, 1999; Craik et al. 1990; Salthouse 1990; Daigneault and Braun 1993), and explicit or elaborative processing (MacDonald, 2008; Stanovich, 2004; Satpute & Lieberman, 2006; Dehaene & Naccache, 2001; Gray, 2004; Koch, 2004; Miller & Cohen, 2001) are reasonably synonymous terms in many ways. There may be debate in regard to specific mechanisms and roles of these different functions but it is generally agreed that these high level processes include; performance monitoring, planning, and strategy selection (Baddeley, 1986), maintenance of goal representations (Cohen et al. 1996; Braver and Cohen 2000; Braver et al. 2002) and inhibition of distracters and automatic processes (Duchek et al. 1998; Friedman and Miyake 2004; Spieler et al. 1996; Zacks et al. 1996; West and Bell 1997; Mayet al. 1999), intentional shifting of attention, and monitoring and revising of working memory (Craik et al. 1990; Salthouse 1990; Daigneault and Braun 1993; Eysenck, Derakshan, Santos and Calvo, 2007, Miyake et al. 2000; Baddeley 1992; Baddeley, 1986; Baddeley, 2001; Smith and Jonides, 1999).

Collectively, executive functions have limited capacity in that they tend to falter following mental exertion, stress, or when cognitive task demands are relatively high (cognitive load e.g., Aspinwall and Taylor, 1997; Baumeister et al., 1998; Muraven and Baumeister, 2000; Trope & Neter, 1994). Likewise, explicit cognitive processes typically inhibit automatic processes and responses to prepotent affective stimuli (“effortful control” MacDonald, 2008; Buss, 2005; Blair & Razza, 2007; Zelazo & Cunningham, 2007; Richeson & Shelton, 2003; Richeson, Trawalter, & Shelton, 2005; Schmeichel & Baumeister, 2004; Ellis & Ashbrook, 1988; Darke, 1988; Forgas, 1995). However, depending upon the strength of affective stimuli and prepotent responses, explicit processing may be compromised. The result of such compromise logically equates to cognitive bias.

Do implicit and explicit processes interact?

Although some researchers believe that implicit and explicit processes are independent, this idea is highly debatable (Batchelder & Riefer, 1990; Bodner, Masson, & Caldwell, 2000; Curran & Hintzman, 1995; Jacoby, 1998; Joordens & Merikle, 1993; McBride, Dosher, & Gage, 2001; Jacoby, 1998; Challis and Brodbeck, 1992, Mathews and MacLeod, 1994; Ashby et al., 1999; Heller & Nitschke, 1997; Tomarken & Keener, 1998). For example, intention can override mood-congruent recall bias (Parrott and Sabini, 1990; Schwartz and Clore, 1983; Mathews and MacLeod, 1994). Indeed, fMRI studies have indicated that whereas interpretive processing mediated by the dorso-parietal cortex (DPC) can be influenced by the stimulus driven processing associated with the ventro-parietal cortex (VPC) when infrequent events occur. Conversely, stimulus driven processing in the VPC can be altered by shifting DPC related goal objectives (memories) (Cabeza, 2008; Corbetta and Shulman, 2002). These ideas are consistent with some models of anxiety-related bias that suggest anxiety may also play an influential role in evaluative and interpretative processes as well as in attention (Mathews and Mackintosh 1998; Mogg and Bradley 1998; Curtis and Locke, 2007). Moreover, processing of affective stimuli may interact with controlled cognitive functions so that demands on one affect the efficacy of the other (Schmeichel, 2007). For example, inhibition of responses to affective stimuli borrows from working memory functions and vice versa, high demands on working memory decrease effective inhibition of responses to affective stimuli (Engle, Conway, Tuholski, & Shisler, 1995; Roberts, Hager, & Heron, 1994). Moreover, completion of tasks that require significant controlled processing appear to subsequently increase the subjects’ vulnerability to emotional stimuli while subjects who first engage in an affective control task are subsequently impaired in working memory tasks (Schmeichel, 2007).

Do different motivational states affect different levels of cognitive processing?

Different affective/motivational states appear to be able to exert cognitive bias in different ways. For example individuals who have diagnosis related to anxiety and individuals that are diagnosed with depression demonstrate related cognitive biases differently (Mathews and Macleod, 1985). Subjects who are prone to anxiety show reliable increases in latency of response to threat-related stimuli in stroop tasks (MacLeod and McLaughlin, 1995; Mathews and MacLeod, 1985; Macleod and Rutherford, 1992), even when the word stimuli are masked and cannot be consciously identified (Harvey, Bryant, & Rapee, 1996; Lundh, Wikstrom, Westerlund, & OÈ st, 1999; MacLeod & Hagan, 1992; MacLeod & Rutherford, 1992; Mogg, Bradley, Millar, & White, 1995; Mogg, Bradley, Williams, & Mathews, 1993a; Mogg, Kentish, & Bradley, 1993b; van den Hout, Tenney, Huygens, & de Jong, 1997; van den Hout, Tenney, Huygens, Merckelbach, & Kindt, 1995). This indicates that the attentional bias for threat-related information is the result of processes which occur before conscious processing. Clinically depressed patients also demonstrate cognitive bias for negative word stimuli ( Dobson and Shaw, 1987; Bradley and Mathews, 1983; Mathews and Bradley, 1983; McDowell, 1984; Clark and Teasdale, 1985). In contrast to anxiety related bias, cognitive bias for mood-congruent stimuli in depression appears mostly in explicit tests of memory that infer elaborative processing ( Mathews and MacLeod, 1985).

It is possible that differential expression of bias at different points in processing may reflect characteristics of the related stimulus. Many previous studies have reported that high anxiety subjects demonstrate increased engagement by threat stimuli and low anxiety subjects demonstrate avoidance (disengagement) of the same stimuli. This has been argued to mean that anxious individuals may have difficulty disengaging from such stimuli while healthy subjects quickly assess threat potential and then immediately disengage as no viable threat is detected (eg..Koster et al., 2006; Fox, Russo, & Dutton, 2002; Yiend & Mathews, 2001; Vasey, El-Hag, and Daleiden, 1996). However, the difference in processing of threat stimuli between high and low anxious subjects may be a function of stimulus intensity (Faunce, 2002; Lee & Shafran, 2004; Williamson, White, York-Crowe, & Stewart, 2004; Dobson and Dozois, 2004; Mogg, et al., 1998).

It is intuitive and logical to imagine that the hungry person finds food stimuli to have increased salience (Mogg, Bradley, Hyare, Lee, 1998), but it is also logical to imagine that the person who is not biologically hungry may find themselves attending to and approaching food stimuli when hedonic qualities of those stimuli are especially intense. Consistent with this idea are observations that flavor strength (Szlyk, Sils, Francesconi, Hubbard, Armstrong,1989) sweetness (Rolls, Fedoroff, Guthrie and Laster, 1990) and temperature (Boulze, Montastruc, Cabanac, 1983) all influence elective ingestive behaviors. In general, stimulus intensity is directly related to efficiency of vigilance task performance (Warm & Jerison, 1984; Matthews et al., 2000; Temple et al. 2000; Warm, 1993; See et al.,1995), an observation that is clearly consistent with the idea that stimulus intensity is a significant factor in controlled cognitive processing and is undoubtedly important in implicit cognitive biases as well.

In regard to threat-related stimuli Wilson and MacLeod (2003) have demonstrated similar intensity related influences. They used an attentional probe task to study bias in high trait anxiety subjects vs. low trait anxiety subjects. In this task they varied the expressions of human faces using a “morph” program to produce differing intensities of threat-like expressions. All subjects showed longer latencies for responding in association with high threat stimuli, but the high trait anxiety patients also had longer latencies in association with stimuli of lower threat intensity. Low trait anxiety subjects in fact displayed faster responses to lower threat stimuli (interpreted as an avoidance response or controlled disengagement). The observation by Wilson and MacLeod (2003) that all subjects show slowed responding to higher intensity threat stimuli suggests that anxiety prone subjects simply have a different threshold sensitivity to threat stimuli and that stimulus intensity is a major influence in bias (Koster et al 2006; Li, Wang, Poliakoff, Luo,). This argument may either call into question many results from past studies where stimulus intensity was not examined, or alternatively may clarify discrepancies that have clouded arguments regarding the nature of cognitive bias in anxiety states. The role of stimulus intensity in the study of other cognitive biases has not been widely assessed (Field and Cox, 2008), but the possibility that argued differences in the influence of different cognitive biases are largely due to intensity factors (or test procedures ; Neill, Beck, Bottalico, and Molloy, 1990), can not be ruled out at this point.

Does cognitive bias affect motivational state?

Williams, Mathews and MacLeod ( 1996) have suggested that a vicious cycle may develop in anxiety related disorders in that initial preoccupation with anxiety (or depression related) related stimuli , leads to attentional “hypervigilance” to cues indicating danger, which in turn increase their sense of anxiety. They further suggest that this leads to a bias in estimation of danger, which is consistent with the exaggerated sense of danger that patients with panic disorder reflect in their interpretation of panic-related body sensations (Clark, 1988) and in measures of sensory sensitivity to relevant stimuli ( Burgess et al., 1981; Foa and McNally, 1986; Powell and Hemsley, 1984). Extending the evidence for this notion, MacLeod, Rutherford, Cambell, Ebsworthy and Holker ( 2002) found that the manipulation of attentional bias to emotional stimuli in dot-probe procedures led to increased negative emotionality in a post dot-probe anagram stress test. This outcome is consistent with other studies (Foa et al., 1993; Hope et al., 1990; Mathews and MacLeod, 1985; McNally et al., 1994; Thrasher et al., 1994). Although appetitive and aversion-related motivations are argued to work differently (Cacioppo and Berntson 1999; Lang et al. 1997), similar effects of attentional manipulation have been reported for appetitive motivations as well (Mogg and Bradley 1998; Gotees, Robinson and Brian 2008). Attentional training could be looked at as priming and it seems very plausible that such effects are implicated in important behavioral problem areas such as the sensitization to drug related cues that is associated with drug addiction (Robinson and Berridge 1993, 2003; Wise 1996; Wiers, vanWoerden, Smulders, & de Jong, 2002).

Cognitive priming and cognitive bias

Studies of automatic responses produced by priming may tell us more about how intentional and automatic processes can interact. Priming studies relate to cognitive bias in that they examine the influence of nonconscious or preconscious priming stimuli on information processing and behavior. Bargh (1989) describes that relevant stimuli automatically capture attention and activate semantically related associations (see Wegner & Bargh, 1998; Hasher and Zacks,1979), that in turn may lead to related changes in cognition and behavior. Priming effects are considered to be automatic because priming stimuli cannot be ignored even when they are task irrelevant may impair task performance (see Perlman & Tzelgov, 2006; Tzelgov, 1997 Forster, Liberman and Friedman, 2007). Priming according to Bargh must involve exposure to either subliminal stimuli or supraliminal stimuli that are not recognized as primes. He has conducted a large number of studies showing that word primes elicit activation of schema (mental concepts) such as the concept of being intelligent, being polite, exerting power, being cooperative and attaining achievement, that exert influence over cognitive processes ( expectations, self-reported information, recall biases), and behavior. For example, Bargh and Pietromonaco (1982) have reported that subjects who had been exposed to subliminal word stimuli that were related to hostility subsequently gave more negative ratings of a standard other person than did those exposed only to neutral control words. Bargh argues that current “goal or motivational state mediates the effects of multiple primes by first filtering through selective attention what is processed and secondly through controlled over-ride of competing or subordinate primes (Barg, 2006; Moskowitz, Gollwitzer, Wasel and Schaal, 1999). Competing goals in this field of research are often referred to as temptations, and controlled override is often referred to as self-regulation (Dhar & Wertenbroch, 2000; Freitas, Liberman, & Higgins, 2002; Gollwitzer, 1999; Kivetz & Simonson, 2002; Kuhl, 1986; Loewenstein, 1996; Metcalfe & Mischel, 1999; Muraven & Baumeister, 2000; Trope & Fishbach, 2000), and self-regulation in general is considered to be ( as described above) a conscious process that requires effort (Gollwitzer,1990; Mischel et al., 1996; Muraven & Baumeister, 2000; (an interesting question here would be to ask if “counter priming” (priming in a direction that conflicts with a goal or innate bias) may modify such biases).

Goal shielding theory; Cognitive bias in goal prioritization?

The override of competing primes need not exclusively be a controlled process. Automatic and potentially implicit prioritization of goals seems to be a plausible explanation for why we are not constantly running to buy every item that we encounter (subliminally or supraliminally) on billboard advertisements and television commercials. According to goal shielding theory (GST; Kruglanski et al, 2002 Shah, Kruglanski and Friedman, 2002) goals exist in a hierarchy with superordinate, subordinate, competitive and complementary characteristics. This theory predicts that goal activation will automatically activate processing of means toward that goal, and will inhibit the processing of stimuli associated with conflicting or competing goals. In a modified lexical decision task (Kruglanski et al, 2002), subjects were subliminally primed with one of three goal attribute words or nonattribute words before conscious processing of target words. Target words related to the subjects (previously self-reported) goal attributes, or were control words. Subjects then made rapid decisions about target words in regard to whether the word represented a personal goal attribute. When subjects were primed with nonattribute control words, response times to target words and non-target words did not differ. However, when primed with one of the subjects’ goal words, words associated with their other goals were processed significantly slower, while non-goal word processing was not affected. Their interpretation was that priming of one goal inhibited the processing of competing goal stimuli (goal shielding). Further, priming with words related to highly committed goals produced greater inhibition in the processing of alternative goal stimuli. Goal shielding seems to occur automatically as it is observed with subthreshold priming of committed goal constructs (Shah, 2002). A number of other studies also support the idea of automatic goal shielding (Goschke and Dreisbach, 2008) and that implicit processing of contextually relevant environmental cues have influential power in self-regulation (Aarts & Dijksterhuis, 2000; Bargh, Gollwitzer, Lee-Chai, Barndollar, & Troetschel, 2001; Kruglanski et al., 2002; Moskowitz etal., 1999; Shah & Kruglanski, 2003).

Obviously, the nature of goal dynamics and interactions must be complex (Fishbach and Zhang, 2008; Ferguson and Bargh, 2004; Fishbach, Shah and Kruglanski, 2004; Moors, De Houwer and Eelen, 2004). It is self-evident that we may have multiple goals operating simultaneously, and that different goals dominant in different contexts. It is also fairly obvious that we do have some degree of conscious control in the choice of goals we strive for, how much time and effort we give to each and decisions of management in the reasonable juggling of simultaneously held goals. For example, when Johnny competes in a soccer match he becomes intensely involved (because he is highly committed) despite the fact that he is fatigued from sleep deprivation (and desires sleep) and is under stress from final exams (and needs to study). He notices a girl in the crowd that he is infatuated with( wants to “date”), hanging out with a potential competitor (he wishes to thwart). In addition, Johnny anticipates that after the game he can eat because he is famished. Yet despite the existence of these simultaneously appreciated goals, he focuses (goal commitment and goal shielding) in the moment necessary to react to his teammate making a pass to him that allows him to skillfully score the winning goal. As time passes, Johnny’s goals and their priorities change. Some goals have been frustrated to the point of abandonment and other goals have met successful conclusions and so have been replaced with other goals. On the other hand, it is equally realistic that we may often be deterred in our goals either by circumstance or in other cases by failure to stick with the goal-related efforts. For example, the next day Johnny attends his lecture classes intending to take serious notes, but falls asleep in the class instead, and in consequence feels guilty and anxious about his grade prospects. Automatic affective influences on goal efforts that lead to interruption, distraction, or failure to maintain goal efforts as well as affective influences on reprioritization and override of our prudent goals are especially relevant to the issue of cognitive bias. There is general accumulating evidence that the dynamics of goal selection, prioritization, shielding and override may be influenced by affective state. (Heller & Nitschke, 1998; Tomarken & Keener, 1998; Ekman & Davidson, 1994; J. A. Gray, 1990; Lang, 1995; Lazarus, 1991; Schwarz, 1990; Simon, 1967).

It appears likely that in some contexts affective or motivational states again act as a steering currents in goal prioritization. For example, goal shielding effects increase with anxiety and decrease with depression (Shah, Friedman and Kruglanski, 2002) indicating an intrinsic relationship between affect and goal prioritization (cognitive bias) (Bargh, 2006; Lerner, Small, Loewenstein, 2004; Carver and Scheier, 1998; Linville, 1996; Neill, 1977; Tipper, 1985; Kuhl and Helle,1986; Kuhl & Beckman, 1994). Other recent studies have shown that specific emotional constructs such as guilt can be unconsciously primed and that this priming automatically leads to logically goal-related behavioral responses (lower self-indulgence and helping; Zemack-Rugar, Bettman and Fitzsimons, 2007; Mauss, Cook, Gross, 2007; see Forster, Liberman and Friedman, 2007 for review). Likewise Ramanathan and Menon (2006) have demonstrated that hedonic goals (temptations) are potent competitors (especially when the hedonic temptations are chronic) to goals of self-control (self-regulation; note that Ramanathan studies did not examine the avoidance end of the hedonic spectrum; nor did the study consider role of social constraints on hedonic temptation i.e. mores, rules, punishments, rewards strategy, negotiation, persuasion etc..). So goals that are so critical to the effects of priming, may themselves be biased by preexisting or co-existing motivations.

One might consider cognitive bias as a consequence or vulnerability to the potential priming effects of hedonic goal stimuli. Similar to the “viscious cycle” concept described in anxious subjects (described above; Williams, Mathews and MacLeod, 1996), such hedonic priming in turn may produce an increased sensitivity to other hedonic related stimuli and to relevant memories due to the spreading activation effect of the priming (Williams,Watts, MacLeod, and Mathews 1988; Graf and Mandler, 1984). Alternatively, cognitive bias may preexist and directly influence which primes and which goals are most effective. Ramanathan and Meenon, (2006), have argued that the degree to which hedonic temptations can override more prudent goals is related to preexisting differences in impulsivity. Although hedonic impulsivity may at some point in learning be altered by hedonic success, it is certainly a construct that logically preexists the laboratory stimuli and tests employed in controlled studies.

The issue of whether self-generated priming can alter the salience of environmental stimuli and thereby affect goal shielding or override the current goal with competing goal behaviors has been addressed by several studies (Jacoby, 1983; Roediger & Blaxton, 1987b; Schacter & Graf, 1989). For example, when subjects were asked to generate visual images of words, an ostensible priming effect subsequently occurred in an implicit memory task (Moskowitz and Roman, 1992). The effect of self-generated priming occurred without awareness or intention (Moskowitz & Uleman, 1987; Newman & Uleman, 1990; Winter & Uleman, 1984; Winter, Uleman, & Cunniff, 1985). Relevant to cognitive bias, self-generated emotion appears to activate brain circuits that are implicated in spontaneous emotional states (Damasio, Grabowski, Bechara, Damasio, Ponto, Parvizi and Hichwa, 2000).

COMMON BIASES

Logical extension of the cognitive bias concept clearly suggests that this phenomena is likely to exist within all humans in varying degrees. Though empirical evidence supporting the concept of cognitive bias has mostly accumulated from studies of individuals that have clinical disturbances in affective/motivational state, there are a substantial number of studies that have demonstrated evidence of bias in non-clinical samples. Where clinical populations may be disproportionately influenced by cognitive bias, it appears that all of us may ultimately be influenced in similar but less intense fashion.

Another logical extension of the cognitive bias concept is that such bias is likely to exist not only for anxiety and depression, but across all logically inferred motivation/affective states. There is growing evidence to support this idea. Although many motivations are idiosyncratic, other motivations may be common to most if not all human beings. For example; while you may or may not be motivated by athletic competition, we are all undoubtedly motivated by hunger and thirst. The testing of cognitive bias in idiosyncratic motivational states will therefore be difficult and time consuming to approach. However, in some circumstances or in certain subpopulations we might expect relatively common motivations or motivational responses relating for example to aggression, anxiety, drug use or concern, academic performance etc..

Anxiety-Related Bias

Although numerous studies have indicated that people with clinical conditions related to anxiety tend to show significantly disproportionate processing of threat related stimuli ( eg.. McNally, 1999; McNally, Otto, Hornig, and Deckersbach, 2001; Lundh, Czyzykow, and O¨ st (1997) this does not preclude that normal subjects may display similar tendencies. To this point, some studies of anxiety related bias have indicated similar bias in normal controls. For example, MacLeod and McLaughlin (2004) assessed explicit vs implicit memory bias in anxious vs control subjects, and reported that both anxiety patients and control subjects showed significant explicit memory bias for threat word recall after tachistoscopic presentation. ( see also MacLeod and Mathews, 1985; Mogg et al., 1989; Matthews and MacLeod, 1985). McNally, Otto, Hornig and Deckersbach (2001) used Jacobys’ Processes dissociation procedure (Jacoby, 1991; Jacoby,Toth, & Yonelinas, 1993; Toth, Reingold, & Jacoby, 1994) to assess the relative contributions of implicit and explicit processing of threat stimuli in panic disordered patients. Using a word-stem completion task with instructions to either use words they had been exposed to or to not use those words, they found no evidence to support disproportional automatic processing for threat words in patients vs healthy control subjects. Importantly, they did observe a higher rate of threat word stem completions in these patients as well as for their control subjects. They suggested that this may indicate self-generated priming for threat stimuli, but also pointed out that this result could occur due to high base rates for completion of threat stem words. It seems plausible that these results could reflect a more or less common bias for threat word processing over processing of neutral or positive control words. As such, it may amount to “cognitive bias” related to a common underlying motivation/affective state. Moreover, according to Mitte’s (2008) meta-analysis of implicit processing of threat related stimuli in anxiety prone individuals, the evidence of an implicit bias in anxious subjects is questioned ( but see Wilson and MacLeod, 2003). However the same analysis indicated that both anxious and normal subjects demonstrate bias for threat stimuli in explicit memory tasks. Finally, (Wilson and MacLeod; 2003; see also Koster et al 2006; Li, Wang, Poliakoff, Luo,) when higher intensity threat stimuli are used, low anxiety subjects as well as high anxietys subjects show slowed stroop responses. This observation supports the idea that cognitive bias for threat is relevant to all. It is logical that all humans are predisposed to preferentially process threat (Windmann and Kru¨ger,1998; Dijksterhuis & Aarts, 2003). The study also reinforces the important need to assess stimulus intensity in studies of bias.

Negativity Bias

A growing number of studies have provided evidence of cognitive bias in both clinical subjects and in normal subjects who have logically inferred motivation/affective states. Mood-induction for example, in normal subjects can produce mood-congruent biases in attention and memory processing (Bower, 1981; Blaney, 1986; Ottati & Isbell,1996; Schwarz & Clore, 1983; Ekhardt and Cohen, 1997; Clore et al., 1994). In addition to evidence for bias in anxiety and negative emotional states (White, 1996; MacLeod, Rutherford, Cambell, Ebsworthy and Holker, 2002; Sigmon, Whitcomb-Smith, Boulard, Pells, Hermann, Edenfield, LaMattina and Schartel, 2006), it seems that people in general give more weight to negative traits than positive traits in impression formation tasks (Anderson, 1965; Peeters & Czapinski, 1990), they dislike losses more than they like equally large gains (Kahneman & Tversky, 1984), and they make more causal attributions for negative events than for positive events (Peeters & Czapinski,1990). This differential emphasis on negative stimuli manifests itself not only in self-report measures but in behavior (Spence & Segner, 1967) as well as in neural responses to positive and negative stimuli (Ito, Larsen, Smith, & Cacioppo, 1998). Negatively valenced stimuli have also been shown to elicit more attention than positive stimuli. Early work in this domain relied on measures such as the amount of time voluntarily allocated to processing positive and negative information (e.g., Fiske, 1980; Graziano, Brothen, & Berscheid, 1980). For example, Graziano et al. (1980) found that, when participants were given the option to hear either positive or negative feedback about themselves, they chose to listen to the negative feedback for a significantly longer amount of time. Pratto and John’s (1991) stroop study showed that negative words had longer color-naming latencies than positive words, suggesting that negative words were automatically drawing more attention than positive words. From these data, Pratto and John (1991) concluded that people are automatically vigilant for negative information in their surroundings (but see Hakan Lab studies). However, negativity bias may be altered with increased demands on controlled resource utilization (Knight, Seymour, Gaunt, Baker, Nesmith and Mather, 2007).

Hunger/food-related bias

Mogg, Bradley, Hyare, Lee, (1998) found that normal hungry subjects demonstrated attentional bias for food-related word stimuli. Similarly, and on a different level of processing Moskowitz (1976) found that pleasantness ratings for sweetened solutions were marked after an overnight fast but were negligible following a satiating glucose load. In eating disordered patients Smeetsa, Roefsa, van Furthb, Jansen, (2008) found evidence for bias and distraction by food related stimuli in a speeded visual search procedure. Bias for food stimuli in eating disordered individuals has also been reported to affect attention, judgment and memory (Faunce, 2002; Lee & Shafran, 2004; Williamson, White, York-Crowe, & Stewart, 2004; Dobson and Dozois, 2004; Mogg, et al., 1998). Rieger et al. (1998) demonstrated that eating disordered patients showed a tendency to direct their attention towards words denoting a large physique and away from words denoting a thin physique. More recently, using a pictorial version of the dot-probe paradigm Shafran, Lee, Cooper, Palmer, and Fairburn (2007) found robust attentional bias for eating and weight-related stimuli in eating disordered patients in comparison to controls, but less consistent effects for body shape-related stimuli.

Aggression and anger-related bias

Dot-probe attention tasks have revealed automatic bias for aggression related stimuli in prison inmates incarcerated for aggression-related offenses, and for undergraduate students high in aggression (Smith and Waterman, 2003; see also Hakan lab studies). Likewise, cognitive bias is also indicated for anger related stimuli following anger inducement (Van Honk, Tuiten, Van den Hout, Putman, De Haan, and Stam,2001; Ekhardt and Cohen, 1997; Bond, Verheyden, Wingrove and Curran, 2004; Berkowitz, 1990; Todorov and Bargh, 2002). Sex bias in sex and violent offenders Sex stroop - Smith and Waterman, 2004

Sex –related Bias

Studies using modified stroop tasks have revealed bias for sex stimuli in normal subjects (Geer and Melton, 1997; Janssen, Everaerd, Spiering and Janssen, 2000; Conaglen, 2004; Spiering, 2002; Smith and Waterman, 2004). Related studies (Gillath, Mikulincer, Birnbaum, 2008) have reported that sexual interest and arousal are associated with motives to form and maintain a close relationship. In five studies, sex-related representations were cognitively primed, either subliminally or supraliminally by exposing participants to erotic words or pictures as compared with neutral words or pictures. The effects of “sexual priming” on the tendencies to initiate and maintain a close relationship were assessed using various cognitive–behavioral and self-report measures. Supporting the hypotheses, subliminal but not supraliminal exposure to sexual primes increased (a) willingness to self-disclose, (b) accessibility of intimacy-related thoughts, (c) willingness to sacrifice for one’s partner, and (d) preference for using positive conflict-resolution strategies. See also: Conaglen, 2004; Geer and McGlone, (romance bias) 1990; and Hakan lab studies)

Smoking-related Bias

Several studies have reported cognitive bias for for smoking related stimuli in cigarette smokers (Bradley and Wright, 2003; Kwak, Duk, Na, Kim, Kim, Lee, 2007 ).

Cannabis-related Bias

Among cannabis users, those with high levels of cannabis craving had a significant attentional bias for cannabis-related words on the visual probe task,for marijuana related stimuli (Field, Mogg and Bradley, 2004; Hakan lab studies) .

Drug-related Bias

Attentional biases for drug related stimuli in drug addicts (Field and Cox, 2008; Bearre, Sturt, Bruce and Jones, 2007; Hoshi, Cohen, Lemanski, Piccini, Bond, Curran, 2007; Waters, Sayette, Paty, Gwaltney and Balabanis, 2003), for alcohol (Sharma, Albery & Cook, 2001) for nicotine (Gross, Jarvik, & Rosenblatt, 1993), for heroin (Franken, Hendriks, Stam & Van den Brink, 2004) and for cocaine (Franken, Kroon & Hendriks, 2000) have all been reported.

Gambling-related Bias

Cognitive bias has been indicated for gambling related stimuli in compulsive gamblers (Diskin & Hodgins, 1999; Kushner, Thurus, Sletten, Frye, Abrams, Van Demark, Maurer, Donahue, 2008; Wagenaar, 1988; McCusker & Gettings, 1997; Boyer and Dickerson, 2003).

BADE

Bias against disconfirmatory evidence in schizophrenia may be related to cognitive bias (Woodward, Moritz, Cuttler and Whitman, 2006).

Pain-related Bias

Evidence from speeded RT tasks supports the existence of cognitive bias for pain related stimuli in patients with chronic pain syndromes (Pincus and Morley, 2001; Stegen, Van Diest, Van de Woestijne, and Van den Bergh, 2001;Keogh, Thompson and Hannent, 2003). When compared to healthy controls, chronic pain patients were more susceptible to interference from pain-related stimuli, but not depressed or affect-related stimuli (Pearce &Morley, 1989). A more recent study failed to replicate these findings (Pincus, Fraser, & Pearce, 1998), and reported that mood, rather than pain, was the major influence on response latencies to the pain stimuli.

Findings from studies which have investigated interpretive biases for pain-related information have been more consistent. Using an ambiguous word-stem completion task, Edwards and Pearce (1994) established that chronic pain patients produced significantly more pain-related word completions than both healthy controls and health professionals. Similarly, when asked to produce spontaneous associations to ambiguous word cues, such as ‘terminal’ and ‘growth’, chronic pain patients made significantly more health-related associations than controls and health professionals (Pincus, Pearce, McClelland, Farley, &Vogel, 1994). Finally, when presented with a list of homophones followed by a free-recall task, pain patients made significantly more negative health-related interpretations than controls (Pincus, Pearce, & Perrot, 1996). These results were independent of mood disturbance, but correlated highly with self-reported pain.

Integrate:

Pain draws attention (Eccleston and Crombez 1999), and several

lines of evidence suggest that pain processing can interfere with

cognitive processes and vice versa. First, several studies have

shown a deficit in cognitive ability in people suffering from

chronic pain (Kewman and others 1991; Park and others 2001;

Apkarian, Sosa, Krauss, and others 2004; Harman and Ruyak

2005). Second, a number of studies have shown that pain

perception can be attenuated by cognitive tasks or other distractions,

although this is somewhat controversial (for review,

see Eccleston 1995). Third, there is some support for the use of

coping strategies that employ distraction therapies for chronic

pain control (e.g., see Astin 2004). Finally, there is considerable

evidence from functional magnetic resonance imaging (fMRI)

(Petrovic and others 2000; Frankenstein and others 2001;

Bantick and others 2002; Remy and others 2003; Seminowicz

and others 2004; Valet and others 2004; Buffington and others

2005; Wiech and others 2005), positron emission tomography

(PET) (Petrovic and others 2000; Wiech and others 2005) and

electroencephalography (EEG) (Lorenz and Bromm 1997; Dick

and others 2003; Babiloni and others 2004; Houlihan and others

2004) studies suggesting that pain- and cognitive-related activity

interacts in the brain, possibly because of a reliance on shared

neural resources. However, it is not known whether these

interactions depend on cognitive load and pain intensity.

4 key experimental

findings: 1) more intense pain-evoked activity was more

sensitive to attenuation by a cognitive task; 2) the greatest

interaction occurred between the higher pain intensity and the

easy task; 3) pain did not affect activity in cognitive-related areas

of activation except when cognitive load is minimal; and 4) 3

response profile types characterized the forebrain responses to

increasing pain intensity or increasing task difficulty, but

modulation effects were not restricted to a particular type.

Taboo-related Bias

See MacKay and Ahmetzanov, (2005)…

Beauty-related Bias

The bias in favor of physically attractive persons is quite robust (Bull &Rumsey, 1988; Hatfield & Sprecher, 1986; Herman, Zanna, & Higgins,1986). Attractiveness biases have been demonstrated in such different realms as teacher judgments of students (Clifford & Walster, 1973), voter preferences for political candidates (Efran & Patterson, 1974), and jury judgments in simulated trials (Efran, 1974).

Body sensation/chronic fatigue-related Bias

Cognitive bias for for body sensations in Chronic Fatigue Syndrome (Hou, Moss-Morris,, Bradley, Peveler and Mogg, 2008) have been reported. Moss-Morris and Petrie (2003) reported that CFS patients have an interpretive bias for somatic information which may play a part in the maintenance of the disorder by heightening patients’ experience of physical symptoms and helping to maintain their negative illness schemas. Although patients did not show an attentional bias in this study, this may be related to the methodology employed. The finding that CFS patients have an interpretive bias for somatic information is consistent with both schema theory (Beck, 1991) and cognitive behavioural models of CFS (Surawyet al., 1995). Previous work has shown that CFS patients have a particularly negative schema of their illness and a fixed belief that their illness is largely physical in nature (Moss-Morris & Petrie, 2001; Weinman et al., 1996). On a fundamental level, words may be inappropriate stimuli to test for an attention bias for somatic stimuli (Pincus & Morley, 2001). Paradigms which test for attentional effects using actual physical sensations or stimuli may better. Although CFS patients did not demonstrate an interference effect on the Stroop task they were significantly slower than the healthy controls to colour name regardless of the content of the stimuli.

Other Potential motivationally linked cognitive biases

defensiveness (Jansson, Lundh, Oldeburg, 2005; Jansson and Lundh, 2005),

paranoia (Peer, Rothmann, Penrod, Penn, Spaulding, 2004),

social norms (O'Gorman, Wilson, Miller,2008; male toughness, Bruch, 2007; O'Gorman, Wilson, Miller, 2008),

confirmational bias in crime investigation (Ask and Granhag, 2005; Baldwin, 1993; Leo, 1996),

GENERAL ISSUES TO EXPLORE

Relationship of goal shielding to cognitive rigidity…to zealousness. To epistemological inflexibility??

Goal Shielding

Self-generated beliefs in self-generated lies

Self-deception, rationalization, Inattentional blindness etc as they relate to goal shielding

Impulse control vs. impulsivity

Cognitive rigidity vs distractibility/ spontaneity

IAB vs suggestability…

It seems logical that inattentional blindness (Mack Rock 1998) may be the logical functional end-point of automatic goal shielding in cases where one is highly committed to a goal (in the case of IAB to short term goals described by the researcher that require significant “load.”) Since it is difficult to accept that laboratory tasks represent especially important subgoals for overarching academic goals, that commitment to such a task may instead relate to obedience to authority or to suggestibility or both. Consequently I wonder if there is a meaningful relationship between suggestibility and IAB.

-Is there a way to get more continuous measures of IAB without involving unnecessary demand characteristics?

-What are the roles of implicit processes in ostensibly explicit/strategic processes?

Even stimuli that are explicitly processed are only partially processed.

-What is the mental state that is objective and intellectual, is it truly free of bias? If it can be is this a personality trait or can it be learned/trained.

-One potential problem that has not been widely appreciated in studies of cognitive bias is the concept of stimulus credibility in regard to the use of word or pictorial stimuli. Although word stimuli do have relevance for concept activation of associative networks (Graf, 1984 Bower), on some level of cognitive assessment, such stimuli must ultimately be deemed inconsequential. Stimuli more clearly attached to the environment may constitute more appropriate choices in our studies. Efforts to utilize realistic potential threat stimuli have been sparse. Word stimuli are also suspect in that their use and interpretation may vary widely across individuals and across time. Therefore it may be useful to examine bias with stimuli in alternate modalities of testing i.e. auditory, pictures, Real stimuli ( odor of cannabis, sounds of thunder or gunshot, sex etc)

-Brain circuits underlining cognitive bias and controlled processes

-brain circuits underlining specific motivational/affective states

FORMALIZED PREDICTIONS COGNITIVE BIAS THEORY

1-Any Common motivation/affect that can be logically identified or inferred will influence cognitive processes. To test this prediction requires testing across a range of logical motivational constructs. Fatigue, Breathing, Thirst, elimination, Social exclusion, social status/subordination, jealousy, romance, social evaluation, Social Desirability, Thermally motivated behavior…

2- manipulation of any affective/motivational state will affect bias at some point in cognitive processing.

3-The level at which a cognitive bias exerts itself in cognitive processing will be affected by stimulus intensity and motivational intensity.

4-Individual motivations may also influence cognitive processes, but the most sensitive tests would be necessary to appreciate such unique motivations.

5-Motivations may influence these processes without conscious awareness

6-Cog bias may be produced/altered by priming, if the prime relates to a preexisting affect in the individual. Priming effectiveness should be related to stimulus credibility the word “threat” vs a real life threat.

7-Cognitive bias may be altered by manipulating cognitive load

Cognitive load; as task demands increase the demand for controlled processing, distractibility increases and presumable the influence of cognitive bias as well. Suppression of cognitive bias may result in rebound when task requirements are less restricting

8-Stimuli associated with cog bias will be processed preferentially and will distract from effective processing of proximal stimuli that are more neutral in nature. Cognitive bias is hypothesized to produce heightened distractibility to task irrelevant stimuli ( MacLeod et al., 2002). For example, accumulating empirical evidence indicates that the performance of anxious individuals is more impaired by distracting stimuli vs. nonanxious individuals (e.g., Calvo & Eysenck, 1996; Eysenck & Graydon, 1989; Hopko, Ashcraft, Gute, Ruggiero, & Lewis, 1998; see Eysenck, 1992, for a review).

9-Cognitive biases arise from either fundamental motivations (related to basic survival and procreation), or secondary (learned and/or contextual motivations). Prioritization of biases must occur in order that some motivational states become subordinate to the priority motivation/affect.

Prioritization is likely given to basic biological survival motivations in most situations, but context may alter these priorities such that the motivation to eat food when hungry may become subordinated to motivations for altruistic behavior in the context that other people around you are more hungry than you are…etc

10-external “priming” may act to alter prioritization of motivations when there are not dominant motivational self-generated primes involved.

-Attentional training might be used to induce biases, or modify existing biases.

-Can we train ourselves to be less influenced by such distractors?

11-Cognitive biases can express themselves as interpretational or judgement biases – (Lundh, WikstroÈ and Westerlund, 2001) or inference bias (Siemer and Reisenzein, 2007), or „projection“ or goals (Kawad, Oettingen, Gollwitzer, and Bargh, 2004)

12-Higher order cognitions such as Belief, logic, ambition, professional pursuits, world views may act as secondary (or tertiary) motivational states and therefore be associated with bias for related stimuli.

NEEDS

-the interest tracker- program that will assess time spent evaluating a variety of visual images of different classes of semantically related stimuli. - picture rating task where time reflects interest/motivational salience? Surprise recognition task?

-Eye-tracker system

-Minimized self-report motivational profiler

and cognitive bias profile task

APPLICATIONS

Clinical diagnostic

Jury selection or training

Predictive profiling

Forensics

Self-awareness and growth

Employment assessment

Identification of prejudice and discrimination

Training for mental focus/ decrease of distractability

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download