A. J. Marr
For much more on the theory of the somatic marker, see my new e-book on the psychology of the internet:
The somatic marker hypothesis is one of the most influential explanations of how covert somatic responses guide decision making or choice. Although somatic markers are purported to derive from learning, the somatic marker has never been examined from the sole perspective of learning. It is the purpose of this article to deduce reliable correlations between the bio-behavioral antecedent and consequential events imputed by the primary exemplar of the somatic marker, namely muscular tension. The viability of the somatic marker of tension as a learned or conditioned response is examined, and correlations between tension and behavior can be used to derive testable hypotheses for behavioral control.
key words: Damasio, somatic marker, autonomic arousal, behavioral, mindfulness, tension, IGT
The concept of the somatic marker is one of the most notable and controversial constructs that explain how somatic states enable and change thought, and in a broader sense, the etiology and function of emotion. Postulated by the neurologist Antonio Damasio (1995), the somatic marker hypothesis states that a state of arousal or a ‘gut feeling’ mediated by the peripheral and central nervous systems often precedes and influences decision making, and that arousal is in turn modulated or conditioned by unspecified learning principles. Inherent in the experience of arousal is that it acts as a conditioned or discriminative stimulus that non verbally connotes the overall goodness or badness of response options and assists reasoning by automatically parsing the response options that will be rationally considered that lead to a specific goal.
A distinct attribute of Damasio’s analysis is that the dependent and independent variables representative of the somatic marker are not behaviorally defined, thus rendering the somatic marker difficult to study as a product of learning. A behavioral event will be defined as a discrete and observable response that is directly initiated and modulated across trials by correlated changes in equally discrete and observable informative or discriminative events. These correlations and the predictions derived from them constitute the subject matter of a learning theory. This article will demonstrate that the somatic marker, although ostensibly controlled by learning, is incoherently defined as a learnable or behavioral event, and can only be systematically described from the perspective of learning theory if its dependent and independent measures are precisely defined. This analysis is informed by observable bio-behavioral processes, yet derives the somatic marker from an operational perspective. That is, although a learning perspective can explain covert somatic behavior as evidenced by the somatic marker through an integration of neurophysiological with behavioral events, it may also be used to derive a simple pragmatic model that isolates the personally accessible and manipulable events that permit covert behavior to be predicted and controlled. Conforming to the epistemological principles of a radical behaviorism, this perspective is informed by neuro-physiological processes, but is validated primarily by its predictive power. Another way of conceiving this is describing the somatic marker not from the vantage of its efficient causes as derived from complex neurophysiological processes, but in terms of teleological or final causes (Rachlin, 2007) that represent simple and discrete behavioral outcomes. In other words, a separate question yet to be answered is not how the somatic marker is instantiated through neuro-physiological processes, but what does the somatic marker in fact do. The virtue of this perspective is that it reduces the number of salient factors in its analysis, and produces easily testable predictions that are of major significance in the self control of emotional states.
The Somatic Marker
This somatic marker hypothesis finds primary support from a now classic experiment surnamed the Iowa Gambling Task, or IGT (Damasio,1995). As described by Tomb et al. (2002), “A subject was presented with four decks of cards. After turning over a card, participants either win or lose varying amounts of play money. Unknown to the participants, picking from two of the decks (‘good’ decks) will result in eventual gain, whereas picking from the other two decks (‘bad’ decks) will result in eventual loss. The task ends after the selection of the 100th card, when most normal individuals have picked more cards from the good than the bad decks. After several rounds of picking cards, it was found that ‘anticipatory ’autonomic arousal, as measured indirectly by the skin conductance response (SCR), was significantly higher for bad decks rather than good.” On the primary basis of this experiment, Damasio deduced that arousal acted to non-consciously alert the individual of the bad deck before its ‘badness’ could be rationally determined.
Arousal is the result of previous socialization or learning, and occurs prior to the conscious consideration of response options to alert one to or pre-determine the ‘goodness’ of a particular response set. In other words, autonomic arousal allows one to make a proper choice between response options prior to their rational consideration. The somatic marker (Damasio, 1995) “forces attention on the negative outcome to which a given action may lead, and functions as an automated alarm signal which says: Beware of danger ahead if you choose an option that leads to this outcome. The signal may lead you to reject, immediately, the negative course of action and thus make you chose among other alternatives. The automated signal protects you against future losses, without further ado, and then allows you to choose from among further alternatives. There is still room for using a cost/benefit analysis and proper deductive competence, but only after the automated step drastically reduces the number of options.”
It must be noted that although primarily justified through the peripheral measures of the
skin conductance response (SCR), somatic markers may not involve the periphery at all, but represent “states of bio-biochemical regulation in structures of the brain stem and hypothalamus” (Damasio, 1995). For example, ‘feelings’ of elation, depression, regret etc. certainly ‘mark’ value, but may do so independent of elevated autonomic arousal. In other words, they are psychologically and physiologically distinct from autonomic arousal. Unfortunately, this expands the dependent measures for the somatic marker to include all neurological events and their affective representations that somatically mark the value of behavior, and renders the somatic marker near equivalent to the concept of emotion. In this broader interpretation, the somatic marker hypothesis is not specific to a single event and a single cause, but is rather a taxonomy for affective events originating in the central and peripheral nervous systems, but with the added proviso that they are a-priori assumed to mark long term value. It follows that autonomic arousal is only a type of somatic marker, although it is its primary exemplar. Since the somatic marker has been justified and criticized through the IGT experiment and the attendant use of the SCR as the dependent measure, for our purposes a learning based examination of the initiating cause of the SCR and autonomic arousal, namely muscular tension, will be used as a suitable proxy for the somatic marker.
Finally, in the IGT experiment, the dependent variable of autonomic arousal was indirectly measured through the skin conductance response, or SCR. Autonomic arousal in turn was initiated through processes of socialization or learning. However, Damasio did not provide a systematic explanation of arousal and its physiological and cognitive antecedents. That is, the dependent measure of autonomic arousal is an ensemble of processes that include muscular tension, accelerated heart rate, increased biochemical activity, etc., yet the somatic marker hypothesis does not impute the systematic relationship of these changes, and it does not describe how those changes may be modulated as behavioral events by information or learning. In other words, although autonomic arousal as a somatic marker is imputed to be under the control of learning, Damasio fails to define what element of autonomic arousal is in fact learned. To ascertain this requires first a clear definition of the dependent and independent measures denoted by Damasio in the IGT experiment.
The Dependent Measure of Tension
Essentially, sympathetic autonomic arousal is elicited through the sustained contraction of high threshold motor units of the striated musculature, as occurs during running or isometric exercise (Saito et al. 1986). But arousal may also be mediated by the sustained contraction of small low threshold motor units of the striated musculature (Mcguigan,1991) that are generally employed in the maintenance of posture, and can be measured directly through EMG (electromyogram) or through indirect measures of autonomic arousal (e.g., skin conductance response or SCR; galvanic skin response or GSR) elicited by tension. The SCR and GSR are used extensively in bio-feedback relaxation therapies to train muscular relaxation (Yucha & Gilbert 2004). Physiologically it has been well demonstrated that muscular tension instigates autonomic arousal (Gellhorn, 1967, 1972, Jacobson, 1970, Malmo, 1975). Through a bi-directional connection between the reticular arousal system and muscle efferents, a dramatic decrease or increase in muscle activity throughout the body can respectively stimulate decreases or increases in sympathetic arousal. This striated muscle position hypothesis (McGuigan, 1993) holds that the critical controlling event for autonomic arousal is covert neuro-muscular activity, and that arousal is an artifact of striated muscle conditioning rather than vice-versa.
It is characteristic of tension induced arousal that it is often sustained in situations that entail continuous choice in workaday contingencies similar to those imposed by the IGT. However, a more appropriate measure for sustained or tonic levels of tension is not the SCR, which measures transient or phasic responses, but the SCL, or skin conductance level.(The SCR indexes the net or phasic change in sweat gland activity in response to a stimulus or event (e.g., presentation of a picture or a sound), whereas the SCL indexes the basal or tonic level of sweat gland activity.) Because the SCR has been used near exclusively in replications of decision making behavior such as the IGT, this has resulted in the most salient and common attributes of tension escaping observation. Specifically, tonic levels of muscular tension are commonly produced under continuous alternative contingencies or choices and have been generally noted to modulate avoidance behavior. For example, continuous decision making between alternative contingencies (e.g. doing housework or minding a child, working or surfing the internet, etc.) is associated with sustained or tonic levels of tension that is painful and modulates not effective choice, but avoidance. Surnamed the ‘Cinderella Effect’ from the fairy tale character who was first to wake and last to sleep(Wursted et al. 1991, 1996; Hagg, 1991; Lundberg, 1999), the continuous activation of low threshold motor units or muscles (also called Type 1, slow twitch, or Cinderella fibers) because of this psycho-social ‘demand’ causes them to eventually fail, and thus recruit other groups of muscles more peripheral to the original group, resulting in pain and exhaustion. In addition, as the name Cinderella suggests, these slow twitch fibers are slow to deactivate, and will continue activated even during subsequent intervals of rest (Lundberg et al, 2002). The aversive result of this long term activation conforms with McEwen’s model of ‘allostatic load’ (1998), which predicts that tension and arousal will be maladaptive when there is an imbalance between activation and rest/recovery. Specifically, continuous low level or ‘slight’ tension results in overexposure to stress hormones, high blood pressure, and resulting mental and physical exhaustion.
If tension initiates and is sustained by autonomic arousal, and if arousal cannot occur without tension (McGuigan, 1991), it follows that arousal cannot be directly derived through any other proximal cause. Specifically, this means that arousal cannot be directly modulated by changes in information, and thus cannot be a behavioral event. A similar reasoning occurs for states of physiological activation that sustain other behavioral events such as running. Thus running is a behavioral event because it is directly modulated by changes in information, but the rapid heart beat and other changes that are concomitant with running are not directly modulated by information, and represent instead S-R mechanisms that are physiologically elicited to sustain the behavior itself.
The Independent Measure of Discrepancy
According to the somatic marker hypothesis, the independent measures for somatic markers are ‘primary and secondary inducers’. “Primary inducers are innate or learned stimuli that cause pleasurable or aversive states. Once present in the immediate environment, they automatically and obligatorily elicit a somatic response” (e.g. changes in the musculoskeletal system). “Secondary inducers, on the other hand, are entities generated by the recall of a personal or hypothetical emotional event, i.e., “thoughts” and “memories” of the primary inducer, which when brought to working memory elicit a somatic state” (Bechara, 2005). In the case of the IGT experiment, these inducers are represented by the positive and negative connotations, both immediate and deferred of surprising or discrepant changes in the value of individual cards. The IGT experiment entailed an observation of a subject’s performance under separate response contingencies, each signified by manipulations (card flipping) of a separate deck of cards. Performance under a primary contingency, namely pulling cards from a single deck, would result in positive and negative discrepant changes (i.e., good cards and bad cards) that are perceived, and through experience, are expected to be perceived.
The role of unpredicted positive changes in discriminative stimuli is reflected in discrepancy theories of reward or reinforcement (Donahoe & Palmer, 1993, Berridge, 2004) that demonstrate how positive unexpected variances in the functional relationship between behavior and reward co-vary with changes in the activity of dopamine neurons, and code the difference between the expected and actual value of outcomes (Schultz, 1998). Unpredicted negative changes or counterfactual changes in the discriminative function of a stimulus (Camille et al. 2004, Zeelenberg, 1998) perform a similar function by also coding a difference between actual and expected values.
Discrepancy may also be perceived from an estimate of the relative or contrasting goodness of a response. Specifically, discrepant positive or negative variance may be perceived while performing under a primary response contingency, but also concurrently with a discrepant or surprising variance revealed by alternative feasible schedules of reward, as illustrated by the concept of behavioral contrast (Flaherty, 1996). This occurs when the goodness or badness of a reward is determined through its contrast to other response options that are revealed prior or subsequently to choice, and represent feasible opportunity losses or gains. The relative goodness of a response option may be a discrete significant event that is consciously perceived through rumination or worry, or it may represent a discrete yet relatively insignificant event that may be consciously or non-consciously perceived, and represents a ‘distractive’ loss. Distraction will be defined as a small or a series of small imminent feasible (i.e. avoidable/attainable) losses or gains, consciously or non-consciously perceived (e.g., moment to moment feasible loss/gain of access to internet or email while at work). In the many situations when a plenitude of options make their goodness difficult or impossible to logically calculate, choosing one option will more likely result in a subsequent interpretation of discrepant or surprising losses or gains when comparatively better or worse choices are revealed in hindsight. For example, choosing to wait in one check out line among several at a grocery store will result in tension when other lines move faster relative to yours, but will be rewarding and tension free if your line is the fastest.
Negative Discrepancy and Tension
As defined by Damasio (1995), the somatic marker represents a somatic alarm bell for ‘danger’ ahead that is elicited as a Pavlovian or S-R response due to primary or secondary inducers. As reflected in his interpretation of the IGT, this element of danger may be defined as a negative overall outcome, and is uniquely signaled by negative surprising events (i.e. bad card choices). But what is the ultimate signal for tension, the negative card choice alone or the negative value of the overall deck imputed by that choice? Empirical evidence strongly suggests the former. This is due to the fact that ultimate danger in many cases may be recognized yet not signal tension, and that negative counterfactual events are correlated with tension regardless of whether the overall outcome is negative. For example, positive events that also signal danger have no correlation with sustained tension. This premise is supported by empirical observations of behavior that result in not intermittent but continuous surprising positive outcomes. For example, schedules of reinforcement that match demand with skill and thus entail continuous positive surprise or discrepancy (Csikszentmihalyi, 1990) in otherwise dangerous situations (e.g., rock climbing, extreme sports) are highly correlated with profound relaxation and low autonomic arousal. As an exception, a high magnitude positive surprise may be associated briefly with elevated SCR that acts as a reflexive startle or orienting response, as when an individual is startled due to the magnitude of surprise (e.g. laughter at a comedy show, joy upon winning the lottery), but this rapidly declines with the habituation of the novel event and is followed by a general reduction of SCR (Lang et al. 1990). On the other hand, reinforcement schedules that entail continuous negative discrepancy (e.g. minding a child, waiting on tables) and do not signal overall danger are associated with elevated autonomic arousal due to tonic levels of tension that is reported as stress or anxiety (Wursted et al. 1991, 1996; Hagg, 1991; Lundberg, 1999). The latter also holds true in situations of behavioral contrast, and particularly when the likelihood of discrepant loss prior or subsequent to choice increases as the number of choices increase or when choices cannot be rationally compared (Marr, 2006) such as in the case of affective vs. rational choices (e.g. comparing the pleasurable value of anticipating eating a donut vs. the rational value of minding a diet). In this latter case, the goodness of a choice is gauged not only because of its intrinsic or long term value, but relative to the affective value of alternative choices. Specifically, positive inducers set up a negative contrast to a reinforcing event by raising the affective salience of low value alternatives. Thus the prospect of eating a donut becomes more salient and its loss more significant contrasted to the value of staying on one’s diet due to the fact that ‘wanting’ it is pleasurable. Similarly, when a positive event (e.g. opening a Christmas present) lies in the immediately future, the present moment becomes a lot less valuable compared to what lies ahead, thus causing tension that prompts one to move faster towards a goal, as one ‘can hardly wait’. This contrast effect explains why positive inducers are often correlated with tension, and why positive inducers that do not have contrasting alternatives (e.g., flow experiences) have no correlation with tension at all.
Finally, as elicited by the behavioral contrast between rational and/or affective events, tension cannot predict overall value, and occurs independently of the overall outcome or value of the primary contingency. For example, choosing one pair of shoes among a myriad options at a clothing store will more likely result in tension upon the subsequent appraisal of options foregone or that will be foregone, and not because of the value of the original purchasing decision. This ‘tyranny of choice’ (Schwartz, 2004) has been repeatedly demonstrated by social psychological observations of the correlation of tension, anxiety, and stress with multiple choices.
The Role of Affect
The observation that the anticipation of negative surprising events or discrepancies is uniquely correlated with tonic (SCL) muscular tension and associated autonomic arousal can also be described if discrepancy and sustained tension (including of course the neuro-physiological systems they modulate) are denoted as not just informative events, but also as an equivalent hedonic or affective events. That is, discrepancy and tension not only denote information about means-ends expectancies but also represent painful or pleasurable affective states that in themselves modulate behavior in ways that may not be coherent with rational behavior. Specifically, the affective aspect of discrepancy and the fact that it signals other somatic responses conforms with Damasio’s concept of the primary and secondary inducers, which are defined as stimuli that unconditionally, or through learning (e.g., conditioning and semantic knowledge), can (consciously or nonconsciously) produce states that are pleasurable or aversive (Bechara et al. 2003), and signals other subsequent somatic responses.
To understand affect, it is important to understand its semantics. It is an unremarkable fact in science that identical processes may have different corresponding and equally valid metaphorical representations (Lakoff & Johnson,1999). Thus aspects of vision may be represented by the activity of rods and cones in the retina and attendant neural processes, or may be more simply and no less accurately denoted as the perception of colors such as red, blue, and green. Similarly, the metaphorical representation of the processes of tension and associated arousal can also be represented through the metaphor of aversive or painful states. This reported ‘feeling’ is defined as affect. More remarkable however is the fact that discrepant events are affective events as well, and are also reported as pleasurable or painful. Unpredicted positive and negative changes (Shepperd & McNulty, 2002) in the discriminative function of a stimulus respectively result in self reports of elation or depression, and the experience of negative discrepant events (e.g. embarrassment, regret) as well as negative affect derived from the peripheral nervous system (e.g. having a tooth pulled) also correlates with the affective event of muscular tension and corresponding approach or avoidance behavior (Mellers & McGraw, 2001, Miller, 1992). Also, as discussed earlier, positive discrepant events that establish a negative contrasting loss are also correlated with muscular tension. In addition, the modeling or ‘visualization’ of future positive or negative discriminative events may elicit an approximation of a response before it’s eliciting stimuli actually occur. The elicitation of the sensory and motor elements of a response by this ‘secondary induction’ or ‘anticipation’ is also called priming (Donahoe & Palmer, 1993). Thus a student may feel euphoric or depressed in anticipation of a respective likelihood of a good or bad failing mark in a class, and psychotherapeutic outcomes rely upon changing patterns of thoughts to those that elicit positively affective emotional primes (e.g. optimistic thinking). Finally, as experience demonstrates, pain or pleasure scales with the relative importance and duration of a counterfactual event or events. Thus, bad news ‘hurts’ more and good news feels better as these events scale in magnitude. Also, as the Cinderella effect demonstrates, a continuous succession of small negative feasible counterfactual events (i.e., distractions) can cause severe tension induced pain and emotional distress as a distractive day at work or minding the children calls to mind.
Negative affect may be generated by the peripheral or central nervous systems, and originates from physiological (pain from muscular tension or physical trauma) or cognitive (negative discrepancy or ‘disappointment’) or distractive causes. However, the function of pain, regardless of its source, is generally regarded as not representational but imperative. That is, the function of pain is not to inherently represent the advisability of specific action plans but rather to interrupt and redirect attention, and impose a new action priority to escape (Eccleston & Crombez, 1999). That is, pain signals the molecular moment to moment goodness of an individual response rather than the molar goodness of a response set. Given this fact, it follows that the pain of tension (as measured indirectly by the SCR and SCL) logically serves the same function, which is not to help parse long term decisions (as per Damasio’s interpretation) but to set a new priority for short term decisions that lead to escape.
Although negative affect or the continuous anticipation of negative affect correlates with the affective state of tension, the experience or continuous anticipation of positive affect does not. Like pain, positive affect connotes molecular or moment to moment goodness rather than the long term or molar value of behavior. And as with pain, the momentary goodness may or may not cohere with the overall goodness of behavior. As mentioned earlier, in ‘flow experiences’ (Csikszentmihalyi, 1990) such as creative behavior that represent a succession of moment to moment positive discrepancies or surprises and attendant pleasurable affect, relaxation or low autonomic arousal is commonly reported. This positive affect gives value to momentary behavior independent of its long term or extrinsic results. Specifically, in humanistic and social psychology, positive affect elicited by moment to moment behavior represents the intrinsically reinforcing or ‘autotelic’ aspect of that behavior, where you do it for the sake of doing it, and has no bearing on the valuation of the rational or ‘extrinsic’ value of behavior and has no correlation with increased tension or anxiety (Deci & Ryan, 1995).
Nonetheless, as experience reveals, positive and negative discrepancies are rarely continuous, and as ‘bad or good news’ can populate any experience, resulting in a conflation of pain, pleasure, and correlating tension. This ‘eustress’ may represent the intermittent affect elicited by bad and good news that follows one’s attention to the varying fortunes of a sports team during a game, or the pleasure that may result from a choice and the pain of the opportunity loss that is perceived prospective to making a choice or momentarily after a choice. One aspect of this is that it is easy to conflate the causal attributes of pain and pleasure, and associate tension induced affect as an intrinsic attribute of the positive attributes of a situation rather than the negative. Thus the common belief, originating with Selye (1980), that tension or stress can be elicited by desirable events to enhance or ‘energize’ performance.
Finally, the motivational systems that mediate discrepancy based affect are different psychologically and neurologically from the cognitive act-outcome expectancies that mediates rational goal making (Berridge, 2001), and therefore it follows that affect does not inherently denote effect, but rather constitutes an unconditioned response to abstract properties of the contingency, or more concisely, is a schedule effect. Thus, affect does not denote long term (or normative) value, but rather short term or moment to moment value. In other words, although positive affect may cohere with and enhance (or somatically mark) the importance of long term goals, this coherence is not implicit in affect but rather is an artifact of how discrepancy or novelty is rendered through the schedule of reinforcement. That is, affect is not rendered by the normative properties of a schedule or its end result, but rather by the abstract properties of a schedule, or how discrepancy or surprise is arranged. For example, piece work or fixed ratio schedules of reinforcement require reward to follow fixed and predictable performance before the onset of reward, and are generally regarded by the subject as boring or emotionally painful. However, if perceived progress under the same contingency is unpredictable, then performance is regarded as interesting and pleasurable, even if the long term results are known by the subject to be negative. For example, a teacher reinforces a child’s progress in reading through unpredicted praise, and the child’s positive feelings are coherent with successfully learning to read. But this reinforcement is due to the teacher making sure that praise and it’s accompanying positive affect is coherent with the long term goal of learning to read. Positive affect thus motivates good behavior because it is designed to do so, and not because it intrinsically denotes the overall correctness of choice. Likewise, a casino operator reinforces a gambler’s activity through a succession of small successive wins (which mitigate the occasional large loss), but also ensures that the gambler’s positive feelings towards gambling in the here and now are dis-coherent with his objective knowledge pf his long term prospects. Similarly, an individual may adjust the difficulty of a game to provide enough positive surprise to make the game pleasing and thus motivating, or he may engineer a similar contingency by proverbially waiting until the last minute to do a task, thus insuring that positive surprise (and affect) will motivate him to get to work on time, complete his taxes, etc. In these cases, affect is designed to occur after the fact or knowledge of the correct behavior. Finally, in a variant of the IGT surnamed the Soochow Gambling Task or SGT,(Lin et al. 2007, Chiu et al. 2008), it was demonstrated that the choice of card deck was determined by the frequency of small positive variations in the reinforcement schedule during the performance, resulting in choice that was independent of the expected overall value of the performance outcome.
Thus it may be hypothesized that regardless of whether it is mediated by the central or peripheral nervous system or is positive or negative in nature, no affective event can independently determine or predict rationally conceived long term or molar value. That is, and contrary to Damasio’s hypothesis, so called gut feelings alone cannot predetermine the adequacy of long term decisions. (As a qualifying note, the concept of ‘gut decisions’ may be taken literally or metaphorically. Indeed, decisions made non-consciously through solely cognitive operations are often termed gut level decisions even when affect is not present.)
Originality and Limitations of the IGT
The IGT was designed to track overt and covert responses under alternative contingencies wherein reinforcement is variable or uncertain. Yet, measuring such choice-choice behavior under uncertainty is not new either in terms of the experimental paradigm or in terms of its dependent measures. Indeed, using the SCR to measure covert responding under similar conditions of choice provided the experimental foundation of the Dollard and Miller theory of anxiety (Miller, 1992). What is new is that the somatic events indirectly measured by the SCR were imputed to modulate the effectiveness of long term choice rather than short term avoidance, and that Damasio renders a sophisticated neurological rather than learning based explanation of the SCR. Another more serious problem is that the dependent measures of the IGT, namely the ‘gut level’ responses, are incomplete both for Damasio’s original experiment and its subsequent replications.
As derived from the IGT, the somatic marker hypothesis imputes that an affective ‘gut level’ somatic response (i.e., tension mediated autonomic arousal) results from the appraisal of surprising counterfactual outcomes or discrepancies that are implicit in choice. However, the immediate perception as well as anticipation (priming) of these outcomes is also affective, and this affect is mediated by the neuro-modulator activity (e.g., activity of midbrain dopamine systems) that modulates choice. Thus not one but two ‘gut level’ affective responses may occur in tandem across performance under the IGT. Specifically, neurally based affect as elicited by the perception of decision-outcome discrepancy or priming effects due to the anticipation of discrepancy is generally not addressed in the experimental measures of the somatic states that are induced by the IGT and its various experimental iterations. This is all the more remarkable since an immediate discrepancy and the priming or anticipation of discrepancy may be presumed to act respectively as primary and secondary inducers of somatic markers (Bechara et. al, 2003). But this is not a limitation of the IGT but of the observational tools that may be applied concurrently with the experiment. As originally implemented, the IGT represents an experimental ‘snapshot’ of the covert and overt behavior concomitant with decision making or choice. However, in studies that replicate the IGT, only a subset use psycho-physiological data (Dunn, 2006), and these data generally use indirect measures of muscular tension as indicated by the SCR. The unique psycho-physiological affective responses due to the actual and anticipated perception of positive or negative discrepancy as a subject chooses cards are not measured or represented as affective states. The IGT was also designed to minimize tonic muscular activation by scheduling choice opportunities to follow the return of autonomic arousal to a base level (Dunn, 2006). This removes from consideration the continuous choice opportunities which are characteristic of workaday activities and the tonic levels of muscular tension that these activities induce. Secondly, because the negative discrepancy following a card choice is of low consequence (loss of play money), tonic muscular activation is also reduced. For example, in similar choice experiments conducted in 1940's utilizing mild electric shock as the consequence of a wrong choice, Neal Miller (1971) noted that high and sustained muscular tension lead to the moment to moment aversion of choice, and utilized the data to derive the Dollard and Miller theory of anxiety, which posited that tension based arousal mediated avoidance behavior. Third, even when tonic muscular tension is measured, muscular load as represented by the Cinderella effect is not captured since the SCL measures basal skin conductance, not the allostatic load due to the sequential recruitment of the musculature. Because of this, the affective (painful) nature of sustained muscular activity is not revealed. Because the affective nature of tonic levels of muscular tension is not considered, the SCL and SCR have been interpreted as performing a signaling function that instigates the parsing of choice rather than an affective function that instigates the avoidance of choice. This represents a serious omission in the experimental investigation of the somatic marker, since all the relevant affective events occurring due to choice are not considered.
Overall, in decision making peripheral and neurologically based affective responses often occur in tandem, yet no experiments have measured their correlation with the experimental contingencies that underlay choice. Thus the covert responses that mediate choice between alternative contingencies are measured separately through measures of the source of peripheral (e.g., muscular tension) based affect (e.g. EKG, SCR), or in-vivo brain imaging (Berridge,2004) that measure the source of neurologically based affect (e.g., activity of dopamine neurons). However, none have measured them together as events that co-vary across time and performance (i.e. through a within group experimental design characteristic of Skinnerian behavior analysis). The question is how covert and overt behavior under the IGT or in a more general sense alternative response contingencies or ‘decision making’ can be explained through an integration of the data provided by both of these experimental perspectives as reflected in the comprehensive perspective of a radical behaviorism.
Validity of the Somatic Marker Hypothesis
The somatic marker theory is attractive because it’s major underlying premise matches the guidance of intuition, namely that gut level or affective responses in themselves mediate the goodness of long term choice. But in this regard it is entirely false. Whether perceived as a positive or negative event, affect only mediates short term approach or avoidance. Secondly, the primary and secondary inducers that instigate somatic events are elicited by abstract rather than normative properties of contingencies. In other words, the discrepancy based affect that induces somatic responses is dependent upon how reinforcement is scheduled to follow performance rather than the fact that it does follow performance. Thus not only does affect not modulate effective long term choice, but it is never initiated by information that is derived from a conscious or non conscious appraisal of long term events.
Ultimately, somatic events do mark and change the course of behavior, and are influenced by learning. Nonetheless, if they can only mediate short term choice, the somatic marker hypothesis as derived from the IGT experiment must be entirely reconceived.
Assembling an Explanation
The lessons of the IGT experiment as informed by data regarding the affective neuroscience of choice are simple. First, the negatively affective (i.e. painful) states of tension and autonomic arousal correlate with and precede or attend the occurrence of a negative discrepancy that also is represented affectively as painful. Secondly, if pain has an imperative and not representational quality, then the somatic marker of tension mediates not effective decision making, but decisions that can facilitate avoidance or escape from the painful entailments of tension based arousal. Third, if tension is behavioral, then it and its opposite state of relaxation is dependent upon the arrangement of information as denoted by response contingencies.
These hypotheses have four major and testable entailments.
1. If anticipated pain elicits tension, then tension will occur regardless of the source of that pain.
Pain may originate with abstract information that is processed cognitively by cortically instantiated processes (perception, or thinking) or by input from the peripheral nervous system. For example, an individual may become tense in anticipation of finding out a likely poor course grade, or in anticipation of having a tooth pulled. Both are painful, and as common experience demonstrates, elicit tension based arousal when they are anticipated.
2 . If avoidance or escape cannot occur or will certainly occur, tension will not occur.
If avoidance or escape is perceived as impossible or impractical, loss will still occur but will be rendered non-feasible and will not correlate with tension. Non-feasible losses are defined as imminent opportunity losses or beneficial response options or choices that by habit or circumstance cannot respectively be avoided or performed. For example, while at work we possess the option to leave to see a movie, fly to Paris, or go fishing, but habit or circumstance leads us to be unconcerned about their loss, or feasibility. However, when we are planning a vacation when a choice between such outcomes is incumbent, then choosing between similar response options does represent a cause for feasible loss and tension. Similarly, as perceived imminent feasible loss becomes non feasible due to experience or ‘extinction’, tension will correspondingly decrease. Thus a student who is experiencing test anxiety becomes less anxious if avoidance is eliminated through a ‘time out’ (Gresham & Kern, 2004) that extinguishes tension by reducing the prospect of avoiding prospective loss. Although the student may still experience loss and the regret it entails, muscular tension declines when the student realizes it cannot be reinforced. Similarly, individuals who perceive the likelihood of a significant loss of life or property (e.g. a hurricane or disease) will feel tense when there is still a prospect that the loss will be avoided, and merely regretful or depressed when they know it cannot. Also, when avoidance from physical pain is impossible, as in the case of ‘learned helplessness’ (Seligman, 1975; Gatchel et al. 1977), low autonomic arousal is characteristic.
Finally, when avoidance is certain, tension will also not occur. For example, whereas an inescapable electric shock renders the subject (in this case an experimental animal such as a dog) (Seligman, 1975) ‘helpless’ and relatively anxiety free, completely avoidable painful events (e.g. avoiding a dangling power line seen from a distance) are similarly tension free. Similarly, game animals are nonchalant when they knowingly keep a safe distance from predators that are seen and are thus predictable in the distance.
3. Tension or relaxation occur independent of the level of task demand.
In commonly accepted explanations for tension or stress (Selye, 1980), the source of stress is any type of demand represented by not physical but cognitive operations. That is, much as physical demands (e.g. running, exercising) cause ‘wear and tear’ on the body, cognitive demands (continuous cognitive appraisal or thinking) cause a similar wear and tear represent by systemic autonomic arousal or ‘stress’. Notably, this has been implicitly challenged by well known and amply tested relaxation protocols such as mindfulness meditation (Kabat-Zinn, 1993) that presume that relaxation is induced by a the restructuring rather than avoidance of ruminative thought (e.g. a thought as just a thought).
Moreover, high cognitive demand as represented by complex cognition often occurs with relaxation, and relaxation is necessary to optimize cognition. Indeed, effective thinking more often has a negative rather than positive correlation with tension. For example, declarative reasoning is enhanced during states of rest (Grecius et al. , 2003, Raichle et al., 2001), thus leading to the opposite conclusion that tension is deleterious to effective thinking. Proprioceptive feedback from the musculature is not required for problem solving (Rolls, 1999, Taub et al. 1966, Teuber, 1972), and executive attention and conscious feeling are not interdependent but are dissociable events (Naccache et al. 2004). Finally, patients with pure autonomic failure who cannot generate integrated peripheral responses reported unimpaired motivational decision making, emotional processing, and social cognition (Heims, 2004).
In a critical overview of the relevant literature on Damasio’s hypothesis, Dunn concluded (2006) that peripheral feedback and associated arousal does not reliably influence decision making accuracy, and may reflect the end product of decision making rather than a key feature in its development. Finally, although it is commonly expected that tension will occur in high demand cognitive activities, this is not the case when feasible loss is absent. During high demand activities (e.g. creating art, climbing mountains) that perfectly match with skill and hence entail no feasible loss, low arousal or muscular relaxation is commonly reported (Perry, 1997).
4. Arousal is not unitary.
Arousal is a theoretically incoherent term, as it may apply to distinctive neuro-physiological processes (Robbins, 1997) that have equally distinctive causes. Yet, in contemporary accounts of performance motivation and arousal (Easterwood, 1959, Damasio, 1995), and indeed all contemporary commentary and experiment on the somatic marker hypothesis, these processes are not separated or controlled for, resulting in the confounding of their differing respective causes and effects. Thus, although neurologic arousal as marked by heightened alertness may or may not occur concurrently with the somatic response of tension and autonomic arousal, both are commonly combined and used as a practically indivisible state of arousal. For example, attentive arousal felt when watching a football game or riding a roller coaster often occurs concurrently with tension based arousal, forming a species of stress or ‘eustress’ from the conflation of two psychologically and physiologically distinct events. Similarly, an individual may compare multiple positive surprising outcomes, and may be tense (or excited) due to the contrasting choices, yet still be positively aroused due to the positive outcome itself. Thus, his subjective excitement or ‘joy’ describes two separate arousal states that are induced by two separate causes. Distractive events may also represent eustress, as a worker may be positively aroused by choosing to view his email, and concurrently tense due to the concurrent loss of productivity. Indeed, the classic inverted U curve (Yerkes & Dodson, 1908, Easterbrook, 1959) correlating the relationship between performance and arousal is better explained through the recognition that arousal in incentive motivation may represent not one but two separate dependent variables that co-vary with different independent causes.
For example, if task performance is plotted across the level of demand, the likelihood of moment to moment discrepant gain would at first offset the likelihood of loss, a circumstance that would reverse itself as demand increases. Thus, as a function of increasing demand, alertness and performance increases, but eventually decreases with a corresponding rise in tension due to an increased rate of ‘bad news’ implicated by high demand. Yet high demand does not necessarily implicate difficult or unattainable performance. Indeed, as earlier argued, remove the negative but not the positive discrepancy entailed by moment to moment performance and retain the demand for performance, and tension and autonomic arousal will not increase with demand, and high demand will correlate with a steady state of relaxation and low autonomic arousal. In addition, performance will not trail off and deteriorate with demand, but increase, and corresponding affect will be positive and quite pleasurable. This resulting ‘flow’ experience (Csikszentmihalyi, 1990) thus demonstrates how escalating demand increases alertness and performance, yet when negative discrepancy is eliminated, increasing demand will not result in increased autonomic arousal (or stress) and decrements in performance as the Yerkes-Dodson model predicts.
A New Synthesis
In sum, tension induced autonomic arousal occurs to somatically mark future uncertain and imminently avoidable negative affective events that in turn originate from the peripheral (pain, as in dental office) or central nervous systems (cognitive discrepancy, such as losing or anticipating losing one’s job, or consistent distractions at work). Because it is affective, tension occurs and is reinforced by any decision that results in avoidance of future pain or escape from present pain, and not decisions made by other normative or rational criteria. That is, as painful events, tension and arousal are not representational, but imperative. Tension marks the anticipated goodness of moment to moment responding rather than long term ends.
Unlike priming responses (e.g. positive affect due to thinking optimistic thoughts) which are elicited independently of action (or subsequent reinforcement), the affective response of tension is emitted because it is reinforced by action (namely successful avoidance), and thus is a conditioned response, or is behavioral. Relaxation is thus predicted to be coextensive with the consistent avoidance of feasible loss. Finally, although tension occurs concomitant with choice, periodic choices (e.g. choosing an item from a dinner menu), allow the autonomic nervous system to gradually return to a homeostatic resting state. However, if choice occurs from moment to moment (e.g. continuous choosing during the working day between surfing the internet or working), then as the Cinderella effect predicts, tension will be tonically sustained and result in the physiological and emotional concomitants of stress. In other words, although phasic neuro-muscular activity occurs with periodic choice, tonic neuro-muscular activity is a function of continuous or moment to moment choice; and it is the latter that is commonly associated with the classic symptoms of stress.
This position that tension mediates avoidance conforms with Zajonc’s (1998) argument that emotions are designed to help individuals make approach-avoidance distinctions, and more specifically, that tension based arousal or ‘anxiety’ (Mowrer, 1939, Miller, 1992) is reinforced through its mediation of the avoidance of future painful events. This position contrasts with the true-false determinations that are mediated by Damasio’s somatic marker (Damasio, 1995). Finally, the fact that tension embodies an affective quality distinguishes the hypothesis from the computational role assigned by Damasio. That is, tension does not mark value because it is perceived, but rather it marks value because it is perceived to hurt. This affective value of tension and tension induced arousal serves as a mechanism to interrupt and redirect cognitive processing by bypassing rather than enhancing cognitive filtering and thus expediting action leading to avoidance or escape (Lowenstein et al., 2001, Armony et al., 1997). It is through this facilitation of action that tension is reinforced.
The presumption that tension is instigated by choice and is reinforced by successful avoidance also conforms with social psychological findings regarding the rationality of decision making. As commonly conceived, tension based arousal is mediated by stimulus-response (S-R) processes, and tension respectively diminishes with successful avoidance (i.e. flight or fight) (Seyle, 1980) or when decisions are successfully parsed or made (Damasio, 1995). But social psychological data do not support either of these positions. In fact, the experience of tension based arousal is mediated by future alternative outcomes or choice, not demand. Moreover, tension induced by multiple decisions or choices mediates not effective choice but the avoidance of choice, and does not serve formal principles of rationality (Keys & Schwartz, 2007). It thus follows that since choice avoidance is modulated by contingent future outcomes or reinforcers, tension must represent not an S-R response, but an R-S or operant behavior.
As a final note, it must be remarked that this conceptualization is not new, and returns the analysis of tension to the seminal concepts of Sigmund Freud. Specifically, Freud’s concept of ‘signal anxiety’ (Wong, 1999) posited like Damasio that tension based arousal unconsciously signaled ‘danger’ ahead. But in contrast to Damasio, the results do not make for better choices, but are a cause for avoidance of overt or covert (i.e., repressing thoughts about an object) behavior. That is, tension results in the avoidance of the momentary pain of tension rather than informing the goodness of long term behavior. Subsequently, in a series of experiments involving choice utilizing the SCR as the dependent measure for tension, Neal Miller (1992) reframed the Freudian role of anxiety within a learning perspective, and the resulting Dollard and Miller hypothesis for anxiety posited that the affective (i.e. painful) event of tension occurs in anticipation of physically or emotionally painful events and is reinforced by successful avoidance. Ironically, the Dollard and Miller hypothesis did not gain prominence in large measure because of its simplicity (Bolles, 1994), a salient characteristic that it shares with the equally simple but much more notable hypothesis of the somatic marker. Finally, as with Damasio’s hypothesis, a learning based explanation of muscular tension conforms with the modern widely accepted generalized and updated versions of the James-Lange hypothesis (James, 1894), which propose that bodily signals bi-directionally interact with other forms of cortically instantiated information rather than the former being the determining factor (Dalgleish, 2004).
Tension and Stress
As derived from the IGT experiment, the concept of tension as a somatic marker forces us to rethink the nature of the covert state of muscular tension as not a reflexive but a behavioral or conditioned response. This statement is not necessarily in contradiction to the somatic marker hypothesis itself, which predicates that learning is strongly implicated in the display of covert somatic states. Nonetheless, inferred neurological events are burdened by Damasio with nearly the entire explanation for the somatic marker. This approach, common among psychologists, “limits the possibilities for prediction and control by foreclosing on the pragmatically balanced analysis of how environment circumstances contribute to a given behavioral event” (Moore, 2007). In the case of the somatic marker, this means that the relationship of environmental contingencies to the display of covert somatic behavior is unexplored in deference to the construction of complex neurological models engaging different organelles in the brain. Although in fairness, this may be due in part to Damasio’s core hypothesis that the somatic marker is due to temporally distant outcomes in the future and past that cannot be systematically examined or controlled.
Yet, in contradistinction to Damasio’s hypothesis, tension and autonomic arousal change the value not of long term but of moment to moment responding, which may or may not cohere with rational long term goals. This makes the theory eminently testable under a contingency analysis since tension is originated and controlled by immediately contingent outcomes. Thus, by attributing tension and accompanying arousal to the perception of feasible and affective losses both large and small (i.e. distractive losses), tension becomes a function not of demand but of the feasibility of the avoidance of discrete stimulus events. Thus to be relaxed requires a complete revision of what relaxation entails and its opposite state of tension based arousal or stress.
In his class, the psychologist F. J. McGuigan (1993) would induce relaxation in his students through the technique of progressive relaxation. He would then drop a book to demonstrate how the startle reflex and related tension and associated arousal is inhibited or impossible without the presence of muscular tonus, a finding originally made by Sherrington (1909) and explained neurologically by Gellhorn (1967,1972). However, although all remedies for stress posit relaxation as the necessary antidote for and indeed opposite of stress, the dependent measures for stress in almost all academic opinion represents various measures of arousal as reflected in self reports and physiological and neurological indices (e.g. GSR, EEG). But by confusing the myriad subjective and objective symptoms of stress with its uniform neuro-muscular source, the concept of stress is rendered theoretically incoherent, since there is no one consistent dependent or independent measure for stress that all parties can agree. The result is that stress becomes merely a taxonomy for all non-specific arousal caused by non-specific stimuli categorized by ‘demand’ rather than a clearly defined behavioral response, and renders a panoply of cures that are in some respects worse than the disease.
Before the advent of modern medicine, disease was thought to be caused by an overabundance of impurities or ‘plethoras’ in the blood, and was treated by nostrums such as blood letting that generally caused more harm than good. It was only when illness was redefined by metaphorical representations of discrete viral and bacterial causes that allowed for a scientifically valid therapeutic focus on the prevention and treatment of disease. Similarly, the governing metaphor of tension or stress is that it is a product of an abundance or plethora of demand (Selye, 1980) and that relaxation is a function of reducing the daily demands of life, by means of vacations or pursuing non demanding diversions at home or work. Yet by adding continuous feasible choices in the place of demanding ones, we increase also feasible losses. Thus we retreat from a world of complexity to a world of ever proliferating distractive and worrisome (ruminative) choices, and end up as the Cinderella effect suggests in a state of perpetual tension and emotional exhaustion. Furthermore, by eliminating all demand rather than controlling for aspects of demand, relaxation becomes a mere product of physically and mentally escaping from our troubles, and not an accompanying trait of an active and demanding lifestyle. Specifically, for the day to day stresses created by a world of boundless choices, relaxation is not just a function of radically reducing or restructuring ruminative thought (e.g. meditation, mindfulness mediation), but of radically reducing distraction. That is, because of the ubiquitous choices in the modern world, radically reducing distractive judgements or choices is not incidental to the induction of relaxation, it is fundamental.
This last proscription is novel theoretically, but as an effective therapeutic practice it is ironically commonplace. Because tonic neuro-muscular activity engages a specific class of muscular fibers that are activated easily by even slight distraction and decrease but gradually when eliciting stimuli are removed, the most effective therapeutic interventions that can reduce and eliminate tension must continuously eliminate not rumination but distraction. This intervention is a major component in meditative practices such as mindfulness meditation, the relaxation response, and concentrative meditation, wherein distraction is implicitly eliminated across significant intervals of time. However, the role of distraction has never been explicitly controlled for in meditation research, which has attributed the reduction of tension to the manipulation of conscious rumination (mindfulness meditation) or through obscure S-R mechanisms entailed through attentional focus (concentrative meditation, relaxation response). Indeed, only recently have studies begun to investigate feasible (i.e., distractive loss, rumination) loss as a separate qualitative aspect of cognition that may be used as an independent measure of tension and relaxation rather than task demand. These preliminary findings have lead to the observation that the ability to reduce distraction is a ‘unique mechanism by which mindfulness meditation reduces distress’ (Jain et al., 2007).
Finally, the methodology of learning theory (as exemplified by experimental designs such as the IGT) allows for the analysis of the fine grain co-variations between stimuli and behavior for single subjects mapped across trials (i.e. a ‘within’ group experimental design), and can uniquely demonstrate how practice variables or experience modulate tension and predict its future occurrence. Such co-variations have generally not been considered as independent measures of tension either experimentally or descriptively, and are subsumed under molar concepts such as ‘attention’ or ‘demand’ that can only be imperfectly described through the use of literal language. This is reflected in the many disparate and conflicting definitions of stress. Learning theory mitigates this problem by establishing a much greater precision in the description of the independent variables that correlate with tension and stress. Specifically, if the fine grain or molecular metaphor of ‘choice’ replaces the molar metaphor of ‘demand’ as the primary descriptor of the etiology of tension, then the management of simple contingencies of reinforcement are implicated in a much more precise and uniform description of the origin, prediction, and control of the operational measures of everyday stress. But this avenue has scarcely been researched due in no small measure to the rejection or neglect of the experimental methodology of learning theory as a necessary tool for the analysis of stress. Nonetheless, this argument is won not by the parsimony and precision of a learning based explanation, but through the power of procedure to effect behavioral change. That of course is the mandate and justification of a true science of behavior.
Armony, J. L., Servan-Schreiber, D. Cohen, J. D. & LeDoux, J. E. (1997) Computational modeling of emotion: Explorations through the anatomy and physiology of fear conditioning. Trends in Cognitive Sciences. I. 28-34
Bechara, A, Damasio,H. & Damasio, A. (2003) Role of the amygdala in decision making.
Annals of the New York Academy of Science, 985: 356-69.
Bechara, A. & Damasio, A. (2005) The Somatic marker hypothesis, a neural theory of economic decision. Games & Economic Behavior, 52, 336-372
Bolles, R. (1994) The Story of Psychology: A Thematic History. Belmont, Ca.: Brooks/Cole
Berridge, K. (2004) Motivational concepts in behavioral neuroscience. Physiology and Behavior, 81, 179-209
Berridge, K. (2001). Reward learning: reinforcement, incentives, and expectations, The Psychology of Learning and Motivation, 3, Academic Press, New York.
Camille, N. Coricelli, G. Sallet, J. Pradat-Diehl, P., Duhamel, J. & Sirigu, A. (2004) The involvement of the orbitofrontal cortex in the experience of regret. Science, 304(5674), 1167-1170
Chiu, Y., Lin, C., Huang, J., Lin, S, Lee, P., & Hsieh, J. (2008) Immediate gain is long term loss. Are there foresighted decision makers in the Iowa Gambling Task? Behavioral and Brain Functions, 4(13) (online)
Csikszentmihalyi, M. (1990) Flow, the psychology of optimal experience. New York: Harper Collins.
Dagleish, T. (2004) The emotional brain. Nature Neuroscience, (7), 582-589
Damasio, A. (1995) Descartes Error: Emotion, Reason, and the Human Brain. Avon: New York
Deci, E. C., & Ryan, R. M. (1995) Intrinsic motivation and self determination in behavior, New York: Springer
Donahoe, J.W., & Palmer, D. C. (1993) Learning and Complex Behavior. Needham Heights, Ma: Allyn and Bacon
Dunn, B. D., Dalgleish, T., & Lawrence, A.D. (2006) The somatic marker hypothesis: A critical evaluation. Neuroscience and Biobehavioral Reviews, 30(2): 239-271
Easterbrook, J. A. (1959) The Effect of Emotion on Cue Utilization and the Organization of Behavior. Psychological Review, 66, 183-201
Eccleston, C. & Crombez, G. (1999) Pain demands attention: a cognitive-affective model of the interruptive function of pain. Psychological Bulletin, 125(3): 356-366
Flaherty, C. (1996) Incentive Relativity. New York: Cambridge
Gatchel, R. J., McKinney, M.E., Koebernick, L. F. (1977) Learned helplessness, depression, and physiological responding. Psychophysiology, 14 (1), 25–31.
Gellhorn, E. (1967) Principles of autonomic-somatic integration. Minneapolis: University of Minnesota Press
Gellhorn, E. & Kiely, W. F. (1972) Mystical states of consciousness: Neurophysiological and clinical aspects. Journal of Nervous and Mental Disease, 154, 399-405
Greicius, M. D. Krasnow, B., Reiss, A. L., Menon, V. (2003)Functional connectivity in the resting brain: A network analysis of the default mode hypothesis. Proceedings of the National Academy of Science, 100, 253-258
Gresham, F. G., Kern, L. (2004) Internalizing Behavior Problems in Children and Adults. In Rutherford, R. B., Quinn, M. M., Mathur, S. R. (Eds.), Handbook of Research in Emotional and Behavioral Disorders. New York: Guilford
Hagg, G. (1991). Static Work loads and occupational myalgia- a new explanation model. In P. A. Anderson, D. J. Hobart, and J. V. Danhoff (Eds.). Electromyographical Kinesiology (pp. 141-144). Elsevier Science Publishers, P. V.
Heims, H.C., Critchley, H.D. Dolan, R. Mathias, C.J. & Cipolotti, L. (2004) Social and motivational functioning is not critically dependent on feedback of autonomic responses: neuropsychological evidence from patients with pure autonomic failure. Neuropsychologia, 42 (14), 1979-1988
Jacobson, E. (1970) Modern treatment of tense patients. Springfield, Il: Charles C. Thomas.
Jain, S., Shapiro, S. L., Swanick, S., Roesch, S. C., Mills, P. J., Bell, I., Schwartz, G.E.R., (2007) A Randomized Controlled Trial of Mindfulness Meditation Versus Relaxation Training: Effects on Distress, Positive States of Mind, Rumination, and Distraction. Annals of Behavioral Medicine, 33(1), 11-21
James, W. (1884) What is emotion? Mind, 9, 188-205
Kabat-Zinn, J. (1993) Mindfulness Meditation: Health Benefits of an Ancient Buddhist Practice. In Goleman, D. and Gurin, J. (eds.) Mind/Body Medicine, Consumer Reports Books, Yonkers, Ny.
Keys, D. J. & Schwartz, B. (2007) ‘Leaky Rationality’ How research on behavioral decision making challenges normative standards of rationality. Perspectives on Psychological Science. 2(2), 162-180
Lakoff, G. & Johnson, M. (1999) Philosophy in the Flesh: The Embodied Mind and Its Challenge to Western Thought, Basic Books
Lang, P. J., Bradley, M. M., Cuthbert, B. N. (1990) Emotion, attention, and the startle reflex. Psychological Review, 97(3), 377-395
Lin, C., Chiu, Y. Lee, P. & Hsieh, J. (2007) Is deck B a disadvantageous deck in the Iowa Gambling Task? Behavioral and Brain Functions, 3(16)(online)
Lowenstein, G. F., Weber, E. U., Hsee, Christopher, K. , & Welch, N.(2001) Risk as feelings. Psychological Bulletin, 127(2), 267-286
Lundberg, U. (1999) Stress Responses in Low-Status Jobs and Their Relationship to Health Risks: Musculoskeletal Disorders. Annals of the New York Academy of Sciences, 896, 162-172.
Lundberg, U., Forsman, M., Zachau, G., Eklo"f, M., Palmerud, G., Melin, B., & Kadefors, R. (2002). Effects of experimentally induced mental and physical stress on trapezius motor unit recruitment. Work & Stress, 16, 166-170
Malmo, R. B. (1975) On emotions, needs, and our archaic brain. New York: Holt, Reinhart, and Winston.
Marr, A. J. (2006) Relaxation and Muscular Tension: A Bio-behavioristic Explanation, International Journal of Stress Management, 13(2), 131-153
McEwen, B. S. (1998) Stress, adaptation, and disease: allostasis and allostatic load. New England Journal of Medicine, 338, 171-179
McGuigan, F. J. (1991) Calm down: A guide for stress and tension control. (rev. ed.) Dubuque, IA: Kendall/Hunt
McGuigan, F. J. (1993) Biological Psychology: A Cybernetic Science. New York: Prentice Hall.
McGuigan, F. J. & Lehrer, P. (1993) Progressive Relaxation, Origins, Principles, and Clinical Applications. In Paul M. Lehrer (Ed.). Principles and Practice of Stress Management, 2nd ed. Guilford Press
Mellers, B. A., McGraw, A. P. (2001) Anticipated emotions as guides to choice. Current Directions in Psychological Science. 10, 201-214
Miller, N. E. (1971) Selected Papers. Atherton, Chicago
Miller, N. E. (1992) Studies of fear as an acquirable drive: I. Fear as motivation and fear-reduction as reinforcement in the learning of new responses. Journal of Experimental Psychology: General, 121(1):6-11
Moore, J. (2007) Comments on “Intentional Behaviorism”, by G. R. Foxall. Behavior and Philosophy, 35, 113-130
Mowrer, O. H. (1939) A stimulus-response analysis of anxiety and its role as a reinforcing agent. Psychological Review, 46:553-565.
Perry, J. T. (1997) In the Zone. Chicago, Ill.: Contemporary Books
Rachlin, H. (2007) A behavioral science of mental life: comments on Foxall’s “Intentional Behaviorism.” Behavior and Philosophy, 35, 131-138
Raichle, M. E., MacLeod, A. M., Snyder, A. Z., Powers W. J., Gusnard, D. A., Shulman, G. L. (2001) A default mode of brain function. Proceedings of the National Academy of Sciences. 98, 676-682
Robbins, T. W. (1997) Arousal systems and attentional processes. Biological Psychology, 45, 57-71
Rolls, E. (1999) The brain and emotion. Oxford: Oxford University Press
Saito, M., Mano, T., Abe, H., Iwase S. (1986) Responses in muscle sympathetic nerve activity to sustained hand-grips of different tensions in humans. European Journal of Applied Physiology, 55(5), 493-498
Schultz, W. (1998) Predictive reward signal of dopamine neurons. Journal of Neurophysiology, 80, 1, 1-27
Schwartz, B. (2004) The Paradox of Choice: Why More is Less. Hopewell, N. J.: Ecco.
Seligman, M. E. P. (1975) Helplessness: On depression, development and death. Freeman, San Francisco.
Selye, H. (1980) Selye’s Guide to Stress Research, New York: Van Nostrand Reinhold
Sherrington, C. S. (1909) On plastic tonus and proprioceptive reflexes. Quarterly Journal of Experimental Psychology, 2, 109-156
Sheppard, J. A. & McNulty, J. K. (2002) The affective consequences of expected and unexpected outcomes. Psychological Science, 13, 85-88
Taub, E., Ellman, S. J., & Berman, A. J. (1966) Deafferentation in monkeys: effect on conditioned grasp response. Science, 151, 593-594.
Teuber, H. L. (1972) Unity and diversity of frontal lobe functions. Acta Neurobiologiae Experimentalis, 32, 615-656.
Tomb, I, Hauser, M., Deldin, P. & Caramazza, A. (2002) Do somatic markers mediate decisions on the gambling task? Nature Neuroscience, 5, 1103-1104
Wong, P.S. (1999). Anxiety, signal anxiety, and unconscious anticipation: Neuroscientific evidence for an unconscious signal function in humans. Journal of the American Psychoanalytic Association, 47, 817-841
Wursted, M., Bjorklund, R., & Westgaard, R. (1991) Shoulder muscle tension induced by two VDU-based tasks of different complexity. Ergonomics, 23, 1033-1046
Wursted, M., Eken, T., & Westgaard, R. (1996) Activity of single motor units in attention demanding tasks: firing pattern in the human trapezius muscle. European Journal of Applied Physiology, 72, 323-329
Yerkes, R. M. , & Dodson, J. D. (1908) The relation of strength of stimulus to rapidity of habit formation. Journal of Comparative Neurology and Psychology, 18, 459-482
Yucha, C.B., & Gilbert, C. (2004) Evidence Based Practice in Biofeedback and Neurofeedback. Monograph Wheat Ridge, CO: Association for Applied Psychophysiology and Biofeedback
Zajonc, R. (1998) Emotions. In D. Gilbers, S. Fiske, & G. Lindzey (Eds), Handbook of social psychology, New York: Oxford University Press
Zeelenberg, M. , van Dijk, W., van der Plight, J. Manstead, A., van Empelen, P. Reinderman, D. (1998) Emotional reactions to the outcomes of decisions: The role of counterfactual thought in the experience of regret and disappointment. Organizational Behavior and Human Decision Processes, 75, 117-141