Sleep Paralysis Postepisode Distress
Modeling Potential Effects of Episode Characteristics, General Psychological Distress, Beliefs, and Cognitive Style
James Allan Cheyne
Department of Psychology, University of Waterloo, Waterloo, Canada
James Allan Cheyne, Department of Psychology, University of Waterloo, 200 University Avenue West, Waterloo, ON N2L 3G1, Canada E-mail: email@example.com
 Sleep paralysis (SP) is a brief paralysis experienced when falling asleep or waking up. It is often accompanied by vivid imagery and extreme fear. In addition to the fear during episodes, people often report marked distress following episodes.
With the goal of developing an integrative account of SP postepisode distress, we examined the effects of several potential determinants of postepisode distress sampled from diverse domains: characteristics of the SP episodes (reported fear and vividness of experiences during SP and frequency of episodes), psychological distress sensitivity, supernatural beliefs about SP experiences, and cognitive style. All factors made independent contributions to postepisode distress. A conceptually derived path model integrating these separate factors was tested and largely corroborated. An analytic cognitive style had both direct and indirect effects on postepisode distress. Postepisode distress was found to be approximately equally affected by contextual, cognitive, and affective variables.
sleep paralysis hypnagogic experiences psychological distress supernatural beliefs cognitive style
Sleep paralysis (SP) refers to a brief pre- or postdormital paralysis often, but not always, accompanied by vivid sensory and perceptual experiences, including complex hallucinations (called hypnagogic hallucinations when occurring predormitally and hypnopompic when occurring postdormitally) and, almost invariably, intense fear (International Classification of Sleep Disorders, 2005). To our knowledge, there is no evidence that SP episodes indicate or presage psychoses, although cases exist where hallucinations accompanying SP have led to misdiagnosis of psychosis (Douglass, 2003; Powell & Nielsen, 1998; Shapiro & Spitz, 1976). Although SP has long been associated with narcolepsy as a component of the “narcoleptic tetrad,” its occurrence as an isolated condition unrelated to other symptoms of narcolepsy is at least an order of magnitude greater than the incidence of narcolepsy (Cheyne, 2006).
 Despite the general paralysis during SP episodes, there is sparing of eye movements and continued automatic breathing, though the inability to breathe voluntarily may produce feelings of pressure on the chest and suffocation.  SP can range from a once-in-a-lifetime experience to a recurring phenomenon consisting of frequent episodes occurring in bouts with nightly or multiple nightly episodes.
Substantial variance in reports of SP prevalence estimates (5%–60%) likely reflects, in part, differences in survey methods, operational definitions, age and other demographic characteristics of survey participants, and cultural factors (Cheyne, Newby-Clark, & Rueffer, 1999; Cheyne, Rueffer, & Newby-Clarke, 1999; Fukuda, Miyasita, Inugami, & Ishihara, 1987; Fukuda, Ogilvie, Chilcott, Vendittelli, & Takeuchi, 1998; Kotorii et al., 2001; Munezawa et al., 2011; Ohayon, Guilleminault, & Priest, 1999; for a recent review, see Sharpless & Barber, 2011).
SP episodes are frequently accompanied by hypnagogic experiences, which have been found to sort rather consistently into three factors: intruder, incubus, and vestibular-motor experiences (Cheyne, 2003; Cheyne & Girard, 2007; Cheyne, Newby-Clarke, & Reuffer, 1999).  Intruder experiences include a sensed presence in the room—that is, visual, auditory, and tactile sensations consistent with the presence of a threatening intruder in the room.  Incubus experiences consist of feelings of pressure on the chest, breathing difficulties (choking or smothering sensations), pain, and intimations of imminent death.  Vestibular-motor experiences include feelings of floating, flying, falling, or spinning; out-of-body experiences; and autoscopy (seeing one’s own body, typically from an elevated perspective).
Few people experience all of these; rather, most report various combinations of experiences that tend to load differentially on the three factors. The most elaborate intruder and incubus experiences constitute complex scenarios of threat and assault (physical and/or sexual) and are more strongly associated with fear during episodes than vestibular-motor experiences.
In contrast to relatively detailed analyses of the terrifying aspects of the SP episodes themselves, distress outside of episodes has received little consideration. This neglect also stands in contrast to conventional nightmares, given that considerable research has associated nightmares with subsequent negative affect and psychological distress beyond the immediate context of the nightmare itself (Belicki, 1992a, 1992b; Blagrove, Farmer, & Williams, 2004; Brown & Donderi, 1986; Levin & Fireman, 2002a, 2002b; Levin & Nielsen, 2007; Schredl, 2010; Schredl, Landgraf, & Zeiler, 2003; Zadra & Donderi, 2000). Level of waking distress following nightmares may mediate nightmare frequency and overall well-being (Belicki, 1992a,1992b; Blagrove et al., 2004; Levin & Fireman, 2002b). Belicki (1985) noted that nightmare frequency and subsequent distress were only modestly related, though frequency has often been employed as a proxy for distress.
SP postepisode distress has most frequently been discussed in traditional cultures as well as immigrant populations from such cultures. In traditional cultures, the fear during and following episodes is often embedded in dramatic mythic scenarios involving malevolent and dangerous night spirits (Adler, 2011; deJong, 2005; Hinton, Pich, Chhean, Pollack, & McNally, 2005). Thoughts that a malevolent spirit entity is in the process of sucking one’s soul from one’s body must certainly exacerbate fear inherent in the paralysis experience and subsequent rumination (Cheyne, 2002; Hinton, Chhean, & Pollack, 2005).  Indigenous beliefs about the nature of fear itself—for example, that protracted fear increases bodily and spiritual vulnerability—are another likely source of postepisode distress
(Hinton, Chhean, & Pollack, 2005).
Fear During Episodes
Previous surveys report that over 80% of respondents report fear and two thirds rate their fearfulness during episodes near the top of the scales provided (Cheyne & Girard, 2007; Cheyne, Newby-Clarke, & Rueffer, 1999; Parker & Blackmore, 2002; Solomonova et al., 2008). That fear during episodes is intrinsic to the phenomenon itself is consistent with the finding that those who have had many experiences continue to experience the fear at high levels during each episode (Cheyne, 2005). On the face of it, the fear experience is scarcely surprising, given that finding oneself suddenly paralyzed and semiasphyxiated in the dark is an understandably gripping experience. Paralysis itself is inherently terrifying and sometimes experienced as painful, even in therapeutically induced contexts (Whitham et al., 2011). Moreover, intruder and incubus experiences, often interpreted as threat and assault, are more strongly associated with fear than are vestibular-motor experiences (Abrams, Mulligan, Carleton, Asmundson, 2008; Cheyne, 2003; Cheyne & Girard, 2007).  In addition, fear during episodes has been specifically associated with breathing difficulty and chest pressure or tightness, variables central to the incubus factor
(Hinton, Chhean, & Pollack, 2005; Ramsawh, Raffa, White, & Barlow, 2008; Solomonova et al., 2008). Shortness of breath is a classic symptom of both panic attacks and asphyxia trauma related to near drowning, whereas, more generally, respiratory disorders have been suggested as potential origins of panic disorder (Verburg, Griez, Meijer, & Pols, 1995). Hence, respiratory difficulties during SP may aggravate panic proneness.
The frightening imagery accompanying paralysis seems to provide further explanation for the fear during episodes.  Yet, even normal dreams tend to be characterized by a predominance of distressing situations, negative rather than positive emotions, and negative over positive encounters
(Domhoff & Schneider, 2008; Hartmann, Kunzendorf, Rosen, & Grace, 2001; Parker & Blackmore, 2002). As negative affect in dream narratives increases, so do waking distress and psychopathology (Brown & Donderi, 1986; Pesant & Zadra, 2006).  Given the general pervasiveness of negative affect in dreams, it seems likely that fear is an intrinsic feature of REM. Indeed, the underlying neurophysiology of REM suggests that direct activation of midbrain structures likely generates emotion directly in parallel with the imagery
(Braun et al., 1997; Desseilles, Dang-Vu, Sterpenich, & Schwartz, 2011; Maquet et al., 1996; Nofzinger, Mintun, Wiseman, Kupfer, & Moore, 1997; Nofzinger et al., 1998). Thus, the frightening imagery selected by the brain may be guided and biased by direct activation of emotion centers in the brain (Hobson et al., 2000). Nonetheless, the accompanying sensory experiences likely do augment fear via a positive feedback cycle and perhaps enhance later recall of the experiences thereby playing on the minds of the experients and contributing to distress beyond the episode itself.
In the present study, we investigate the associations of SP episode characteristics—that is, frequency, fear, and the intensity of sensory and perceptual experiences on postepisode distress.
General Psychological Distress
Individuals predisposed to general psychological distress may be especially likely to react with greater distress following episodes. Stress is widely reported by experients across several cultures and subcultures to be a suspected cause of their episodes (Bell, Dixie-Bell, & Thompson, 1986; Fukuda et al., 1998; Ness, 1978; Ohaeri, Odejide, Ikuesan, & Adeyemi, 1989; Spanos, McNulty, DuBreuil, Pires, & Burgess, 1995; Wing, Chiu, Leung, & Ng, 1999; Wing, Lee, & Chen, 1994).  Consistent with these personal intuitions, trauma, life stress, panic disorder, and anxiety have also been found to be associated with SP
(Ramsawh et al., 2008; Sharpless et al., 2010). In addition, Simard and Nielsen (2005) found social anxiety associated with presence experiences during SP episodes. Social anxiety includes fear of scrutiny, which is a common characteristic of the sensed presence in that it is often experienced as a monitoring, watching presence (Cheyne, 2001, 2012). Moreover, in a rare study of SP postepisode distress, Solomonova and colleagues reported associations of social anxiety with SP distress.
To follow up on these findings, we included a standard measure of depression, anxiety, and stress: the DASS 21, a brief version of the Depression, Anxiety, and Stress Scales (DASS; Henry & Crawford, 2005; Lovibond & Lovibond, 1995). Scores on these scales do not necessarily reflect clinical levels of distress but rather the range of levels found in the general population. Because our interest is mainly on distress, we note that the correlations among the three subscales are quite high and that the combined DASS scores have been suggested to assess a “general psychological distress” factor (Henry & Crawford, 2005).
Beliefs About the Meaning of SP
SP has been long and widely associated with supernatural or paranormal entities in historical and traditional contexts, as well as modern demonic-possession and alien-abduction contexts (Adler, 1994, 2011; Cheyne, Rueffer, & Newby-Clarke, 1999; Hinton, Chhean, & Pollack, 2005; Hinton, Hufford, & Kirmayer, 2005; Hinton, Pich, et al., 2005; Holden & French, 2002; Hufford, 1976; Kompanje, 2008; McNally & Clancy, 2005; McNally et al., 2004; Ness, 1978; Wing et al., 1999). Although many moderns are likely skeptical of supernatural explanations, even they may employ them in a metaphorical “as if” sense. Others, however, still accept or at least seriously consider more literal supernatural and paranormal interpretations (Cheyne, 2002).
Supernatural attributions were found to be associated with level of fear and anxiety as well as hallucinations and sensory experiences during episodes in a Japanese sample (Arikawa, Templer, Brown, Cannon, & Thomas-Dodson, 1999). More generally, a larger literature on placebo and nocebo effects of beliefs on reactivity to illness and misfortune reveals the significant impact that beliefs can have on the consequences of SP episodes (Adler, 2011). Nocebo effects of beliefs appear to be pervasive. The most extreme case of the negative impact of beliefs on health can be traced to Walter Cannon’s (1942) notion of “Voodoo death,” in which Cannon argued that the belief that one has been put under a “spell” by one’s own or another’s actions can lead to an extreme stress reaction sufficient to result in death as an apparent consequence of a complex combination of psychological, physiological, and social factors. Much skepticism has been directed at Cannon’s claim, not only because the claimed outcome was so extreme, but also because of the more than implied suggestion that such extreme reactions were characteristic of, if not limited to, “primitive” societies. Subsequent research has generally yielded two general results. First, the general claim that beliefs have health consequences has been supported, though seldom approaching anything quite as dramatic as Voodoo death (Adler, 2011). Second, and more important for the sensibilities of anthropologists, similar effects are also observed among citizens of technologically developed societies, the effect of which was duly labeled the nocebo effect (Eaker, Pinsky, & Castelli, 1992; Luparello, Leist, Lourie, & Sweet, 1970; Luparello, Lyons, Bleeker, & McFadden, 1968; Morrow & Dobkin, 1988; Oftedal, Straume, Johnsson, & Stovner, 2007; Verdecchia, Staessen, White, Imai, & O’Brien, 2002). Thus, a Westerner whose death is accelerated knowing that she or he has been diagnosed with a presumed fatal disease is the modern equivalent of an aboriginal in a traditional society who has been put under a fatal spell (Milton, 1973). This interpretation has been applied to SP, whereby traditional beliefs about the meaning of SP among Hmong people exacerbate the intensity of episodes to the extent that, when combined with protracted stress and certain cardiac anomalies, they appear to have contributed to sudden nocturnal deaths within the Hmong immigrant community in the United States (Adler, 1991, 1994). The present study explores milder but possibly more pervasive effects of supernatural beliefs about the meaning of SP on subsequent distress. We therefore constructed a brief seven-item scale regarding supernatural/paranormal causes of SP and related such beliefs to SP postepisode distress for this purpose (see Appendix).
Cognitive Style and Beliefs
Why different people have often radically different beliefs is a complex question implicating cultural background, family and peer influences, temperament, and much else. Recent research has also implicated cognitive factors, particularly cognitive style, in religious and paranormal beliefs (Gervais & Norenzayan, 2012; Pennycook, Cheyne, Seli, Koehler, & Fugelsang, 2012; Shenhav, Rand, & Greene, 2012). People may differ in their beliefs—indeed, entire systems of belief—in part because they think in rather fundamentally different ways or styles. These differences in thinking styles might then have indirect but potentially far-reaching implications for how we react to life experiences. One dimension of thinking styles that has been found to be important for beliefs is that of intuitive versus analytic reasoning. Considerable research has suggested that human thinking implicates two fundamentally different types of cognitive processing: (a) fast and automatic “intuitive” or “heuristic” processes and (b) slow and deliberative “reflective” or “analytic” processes (for reviews, see De Neys, 2006; Evans, 2008; Kahneman, 2003; Stanovich, 2009). More important for present purposes, there is now substantial evidence for individual differences in the propensity to engage analytic processing (Stanovich & West, 1998, 2000, 2008; Toplak, West, & Stanovich, 2011; see also Epstein, Pacini, Denes-Raj, & Heier, 1996). More colloquially, some individuals may typically elect to “go with their gut feelings” in preference to engaging in more rational, logical, and systematic thinking. Thus, although we all engage in both types of thinking and generally agree that each mode has advantages and limitations, some of us tend to favor one and some the other, both in terms of how we value them (Pacini & Epstein, 1999) and, crucially, how we use them (Stanovich, 2009). In the present work, we were therefore interested in how people differ in their analytic styles. To do this, we used some common problem-solving tests that have been used to assess such differences. Two separate tests contribute to a measure of analytic cognitive style. Because these reasoning measures also tap basic cognitive ability, we also used two tests of cognitive ability to control for this potential confound.
Pathways to Distress
Based on the foregoing review and arguments, a theoretical model is proposed with three sets of paths to SP postepisode distress: a cognitive belief path, an experiential path, and an affective disposition path (Fig. 1). The first path (vertical; top-to-bottom, right-hand portion of Fig. 1) proposes a conceptual pathway from cognitive ability through analytic cognitive style to supernatural beliefs about SP to SP distress. The second path (upper horizontal) moves from SP experiences to SP postepisode distress. A third path (lower horizontal) proceeds from general psychological distress (DASS) to SP distress. Under the assumption that supernatural beliefs can affect episode experiences, a path is also hypothesized from supernatural beliefs to SP experiences.
View larger version:
In this page In a new window
Theoretically derived model of paths to sleep paralysis postepisode distress.
The present data are part of an archival data set composed of participants’ responses to an online SP survey between January and December 2009 and follow-up research based on the e-mail addresses of those who were willing to participate further. Participants were contacted in the fall of 2011 and invited to participate in a study of beliefs and SP distress (see Pennycook et al., 2012, Study 2). The present report includes data from 293 people (mean age = 33.64 years, SD = 12.27, range = 16–69; 223 women, 70 men) who answered all questions regarding postepisode distress, relevant episodes, SP belief, analytic cognitive style, and cognitive ability. The DASS was introduced partway through the study; hence, only 180 participants (mean age = 32.70 years, SD = 11.39, range = 16–66; 128 women, 52 men) answered these questions.
Sleep Paralysis Post-Episode Distress Scale
 The Sleep Paralysis Post-Episode Distress Scale (SPPEDS) is based in part on earlier work on (conventional) nightmare distress by Belicki (1985, 1992a, 1992b) and a three-item SP distress scale employed by Solomonova et al. (2008). The scale consists of five questions (Table 1). The first three concern fairly direct effects and include: ruminating about the episode, difficulty falling asleep, and fear of falling asleep. The remaining two refer to potential consequences to well-being (mood and functioning) the next day.
The questions are ordered below in terms of presumed severity:
“When you recover from a sleep paralysis episode, do you find you keep thinking about it and have difficulty putting it out of your mind?” (Ruminating)
“After you awaken from a sleep paralysis episode, do you have difficulty falling back asleep?” (Difficulty falling asleep)
“Are you afraid to fall asleep for fear of having a sleep paralysis episode?” (Fear of falling asleep)
“Do sleep paralysis episodes upset/bother you the next day? (e.g. cranky, edgy, emotional)” (Mood next day)
“Do sleep paralysis episodes impair your functioning the next day? (e.g., interfere with your ability to do your work, interact with family or friends)” (Functioning next day)
View this table:
In this window In a new window
Item and Scale Means for Sleep Paralysis Post-Episode Scale
For each question, participants were asked to indicate, on a 5-point scale, whether they had the postepisode experience described (1 = never, 5 = very often). The mean scores for the five SPPEDS items were converted to POMP (percentage of maximum possible) scores for ease of interpretation. Means, standard deviations, skew, and kurtosis for each question and for the whole scale are presented in Table 1. The rank order of the means tends to corroborate the presumed differences in severity of distress assessed by the questions, with the items rated as less severe receiving higher frequency ratings. Scores for each item and for the overall scale are generally normally distributed. The five questions appear to be internally consistent as indicated by a Cronbach’s α of .77: a reasonable value for such a short scale, with a mean item-total correlation (r) of .54.
SP episode measures
Before being contacted for the present study, all participants had completed the Waterloo Unusual Sleep Experiences Survey online to determine their lifetime frequency of SP episodes: “When falling asleep or waking from sleep, a brief period during which I am unable to move, even though I am awake and conscious of my surroundings.” Participants were also asked about fear intensity during episodes and intensity or vividness of a variety of sensory and motor experiences during episodes. The Waterloo Unusual Sleep Experiences Survey has been used in whole or part in approximately two dozen published studies in several laboratories (e.g., Abrams et al., 2008; Cheyne, 2003; Cheyne, Rueffer, & Newby-Clark, 1999; McNally & Clancy, 2005; Otto et al., 2006). An episodic version of the survey has demonstrated reasonable test-retest reliability of specific items and major factors (Cheyne & Girard, 2007).
In the present study, fear is rated on a 9-point intensity scale, and episode frequency was rated as occurring once in lifetime, several times in lifetime, several times a year, monthly, weekly, or several times a week. Intruder experiences (feelings of a presence; visual, auditory, and tactile hallucinations; plus an item on experiencing the bed covers being pulled) and incubus experiences (breathing difficulty, pressure, pain, thoughts of imminent death) were combined into a nine-item Threat and Assault Experiences Scale (Cronbach’s α = .73). A Vestibular-Motor Experiences Scale consisted of floating, flying, falling, spinning, elevator sensations, out-of-body experiences, and autoscopy (i.e., seeing oneself from an external, elevated perspective, as well as illusory sitting up and walking around). The Vestibular-Motor Experiences Scale had a Cronbach’s α of .83. Intensity for each experience type was rated on a 0 to 7 scale, with zero indicating no experience of that type (for a fuller description of all items and associated factors, see Cheyne, 2005; Cheyne & Girard, 2007; Cheyne, Rueffer, & Newby-Clarke, 1999).
As mentioned, the DASS 21 is a brief version of the DASS, with 21 items and three subscales: depression, anxiety, and stress (Henry & Crawford, 2005). A total score yields a general psychological distress index.
Sleep Paralysis Supernatural Beliefs Scale
The Sleep Paralysis Supernatural Beliefs Scale has seven items asking about different supernatural beliefs about SP. Following an explanatory preamble, “Below are explanations people often give for such experiences,” participants were asked to indicate the likelihood of each suggested explanation for SP (1 = highly unlikely, 5 = highly likely). Item means, standard deviations, item-total correlations, and internal consistency are provided in Table 2.
View this table:
In this window In a new window
Item and Scale Means for Sleep Paralysis Supernatural Beliefs Scale
Analytic cognitive style and cognitive ability
Two measures of analytic cognitive style were employed to provide some generality to our conclusions. The Cognitive Reflection Test (Frederick, 2005) consists of three quasimathematical problems that generate implicit misleading intuitions, for example, “A bat and a ball cost $1.10 in total. The bat costs $1.00 more than the ball. How much does the ball cost?” The Cognitive Reflection Test is considered a particularly robust measure of the tendency to engage analytic reasoning (Toplak et al., 2011).
We also used three base-rate conflict (BRC) problems that contain a conflict between a salient stereotype and probabilistic information (De Neys & Glumicic, 2008):
In a study 1000 people were tested. Among the participants there were 995 nurses and 5 doctors. Jake is a randomly chosen participant of this study.
Jake is 34 years old. He lives in a beautiful home in a posh suburb. He is well spoken and very interested in politics. He invests a lot of time in his career.
What is most likely?
a. Jake is a nurse
b. Jake is a doctor
In such problems there are two conflicting pieces of information: the base-rate probability of group membership (i.e., 99.5% chance that Jake is a nurse) and diagnostic information that cues an intuitive response based on stereotypical beliefs (i.e., that Jake is a doctor). As with the Cognitive Reflection Test, people who engage in analytic processing are less likely to select the intuitive response (Stanovich & West, 2000).
As the analytic cognitive style measures also reflect cognitive ability (Toplak et al., 2011), two measures free of interfering intuitions were used to control for cognitive ability. WordSum is a brief vocabulary test that correlates well with full-scale measures of intelligence (Davis & Smith, 1994; Huang & Hauser, 1998). For this, participants were presented a list of 10 target words in capital letters and asked to choose the option that most closely matches the meaning of the target word. The second cognitive ability measure consisted of three base-rate neutral (BRN) problems that were structurally identical to the BRC problems except that the personality descriptions were not stereotypically associated with either response alternative. Thus, BRN problems assess the basic cognitive ability required to use the base-rate information to solve the problem without the interfering intuitive response. BRC problems, however, were intended to assess the ability to override a conflicting and, in the context of the problem, misleading intuitive response. Thus, performance on BRC problems, while controlling for BRN performance, provides a measure of preference for probability information over the stereotype, independent of the ability to use probability information, and is therefore reflective of an analytic cognitive style (Pennycook et al., 2012). All cognitive variables were converted to POMP scores (Cohen, Cohen, Aiken, & West, 1999), and unweighted means of the analytic cognitive style variable (Cognitive Reflection Test + BRC) and cognitive ability variable (WordSum + BRN) were then computed separately.
SP postepisode distress
Correlates of SP postepisode distress
A summary table of correlations among all variables is provided in Table 3. Individual correlations are discussed in the appropriate sections below. Following discussion of each set of variables, we assess our integrative model as described in the introduction.
View this table:
In this window In a new window
Pearson Product-Moment Correlations
Consistent with previous studies (e.g., Cheyne & Girard, 2007), fear during episodes was significantly correlated with threat and assault experiences but not vestibular-motor experiences (Table 3). Moreover, the fear–threat and assault experiences correlation was significantly greater than the fear–vestibular-motor experiences correlation by a Williams test, t(293) = 4.40, p < .001. Fear, threat and assault experiences, and vestibular-motor experiences, but not frequency of episodes, were significantly positively correlated with the SPPEDS. In multiple regression analysis predicting distress scores from all four episode characteristics, fear during episodes and threat and assault experiences made significant independent contributions; the effect of vestibular-motor experiences was marginal; and frequency of episodes was nonsignificant (Table 4).
View this table:
In this window In a new window
Multiple Regression Analyses Predicting Postepisode Distress From Episode Characteristics
General psychological distress
The depression, anxiety, and stress measures were each significantly positively correlated with the SPPEDS, with numerically equal values; all r values equaled 0.26, and p values were less than < .001. This is consistent with claims that the three scales may all measure general psychological distress (Henry & Crawford, 2005), and this general factor explains the virtually identical correlations of each subscale with the SPPEDS. In the present study, correlations among the three subscales varied from 0.59 to 0.61. The full DASS 21 scale yielded a Cronbach’s α of .92. We therefore computed an overall DASS score and used this for all subsequent analyses (see Table 3). DASS total scores were significantly positively correlated with the SPPEDS and significantly negatively correlated with age.
Supernatural beliefs regarding SP
Supernatural beliefs about SP experiences were significantly positively correlated with postepisode distress as well as threat and assault experiences and vestibular-motor experiences (Table 3). It should be noted that supernatural beliefs about SP were generally not strong. Means for each item were consistently below the midpoint of the scale. Nonetheless, it appears that holding even relatively weak supernatural beliefs about SP is significantly associated with increased postepisode distress.
Cognitive sources of individual differences in beliefs and the SPPEDS
Analytic cognitive style was significantly negatively correlated with supernatural beliefs and with the SPPEDS (Table 3). Cognitive ability was significantly negatively correlated with supernatural beliefs and positively with analytic cognitive style. Analytic cognitive style remained significantly correlated with the SPPEDS, r p = –.20, p < .001, and supernatural beliefs, rp = –.16, p < .006, when controlling for cognitive ability.
An integrative model: Pathways to distress
The path model illustrated in Figure 1 was tested using structural equation modeling in AMOS V.20 (IBM, Armonk, NY). Missing DASS values were replaced using the FIFA feature of AMOS. For SP experiences, we converted fear during episodes, threat and assault experiences, and vestibular-motor experiences scores to standard scores and calculated a mean SP episode experience score.
The theoretically derived model yielded moderately good fit indices, χ2(9) = 18.08, p < .034, χ2/df = 2.01, comparative fit index = 0.94, normed fit index = 0.89, root mean square error of approximation = 0.06. It was evident, however, given the relatively robust correlation of analytic cognitive style with the SPPEDS that the model might be improved by adding a direct path from analytic cognitive style to the SPPEDS (Figure 2). This model yielded very good fit indices, χ2(8) = 7.29, p = .505, χ2/df = 0.91, comparative fit index = 1.00, normed fit index = 0.96, root mean square error of approximation = .00. Moreover, given that this second model was nested within the first model, it was possible to directly compare the models by examining the chi-square differences. This analysis yielded a robust difference, χΔ2(1) = 10.79, p = .001. The modified model represents a significant improvement in fit over the original model, and analytic cognitive style has a significant direct effect on the SPPEDS independent of small but significant mediation via beliefs and SP experiences, standardized coefficient = 0.05, p < .001.
View larger version:
In this page In a new window
Path diagram, modified from theoretical model in Figure 1, with coefficients based on structural equation modeling of predicted causal paths to SP postepisode distress. All coefficients are significant, p < .002.
SP postepisode distress
The distress items sampled a range of distress reactions from rumination to interference with next-day functioning. The item means were ordered consistent with an a priori ranking of presumed severity of distress. All items had substantial positive correlations with the full scale, and the overall internal consistency was reasonable, given the limited number of items. As expected, the postepisode distress was significantly correlated with several features of SP episodes, as well as distress sensitivity, supernatural beliefs about SP, and cognitive ability and style. Each of these effects is discussed below.
Greater fear and more vivid threat and assault experiences reported during episodes were associated with greater postepisode distress. Consistent with Belicki’s (1992b) conclusions regarding conventional nightmares, SP frequency was not a good predictor of SP postepisode distress. Indeed, the near-zero correlation is even less than that typically reported for nightmares using the Belicki scale (e.g., Levin & Fireman, 2002a). Schredl and colleagues (Schredl, 2010; Schredl et al., 2003) have argued that because the Belicki scale rates frequency of distress (as does the present SP distress scale), there is a bias to a positive correlation with frequency of nightmares. The present results suggest that using a frequency of distress scale does not necessarily bias the correlation with frequency of episodes.
Also consistent with earlier research (Cheyne & Girard, 2007), episode frequency was not significantly correlated with fear during episodes, though both were correlated with vividness of sensory experiences, suggesting that the lack of correlation of fear and frequency was likely not a problem of unreliability.  As expected, those who experienced more fear and intensity of sensory and perceptual experiences during episodes reported more postepisode distress.
General psychological distress
 Researchers have linked SP susceptibility to depression and anxiety, including social anxiety as well as panic disorder
(e.g., Bell et al., 1986; Friedman & Paradis, 2002; Ohayon, Zulley, Guilleminault, & Smirne, 1999; Otto et al., 2006; Paradis & Friedman, 2005; Sharpless et al., 2010; Simard & Nielsen, 2005; Solomonova et al., 2008; Szklo-Cox, Young, Finn, & Mignot, 2008; Yeung, Xu, & Chang, 2005) and the overall DASS findings reported here suggest the possibility that it may not be depression specifically but a more general heightened psychological distress. In addition, reported fear during episodes, in contrast to postepisode distress, is at best only weakly influenced by psychological distress, consistent with the hypothesis that fear during the episode is determined directly by the nature of the SP episode itself. Any integrative account of postepisode distress, however, will likely need to take into account individual differences in associated affective states. The robust alpha coefficient across all DASS items found in the present study is clearly consistent with the suggestion that it measures general psychological distress.
Supernatural beliefs about SP
Those reporting some level of supernatural belief regarding the causes of SP reported increased intensity of both fear and threat/assault experiences as well as increased postepisode distress relative to those who reported weaker or no supernatural beliefs regarding SP.  It is not possible to determine whether supernatural beliefs contributed to greater vividness of experiences or whether vividness of experiences makes supernatural beliefs more plausible, though the path model based on the former assumption worked well.
That supernatural beliefs are not significantly correlated with fear during episodes does, however, suggest that supernatural beliefs increase postepisode distress rather than the reverse. This belief-distress direction of causation is also consistent with numerous communications received by the first author expressing considerable relief from, and gratitude for, naturalistic information about SP. That is, a reinforcement of naturalistic beliefs regarding SP appears to be, in itself, beneficial, though perhaps not equally for all. The direction of causality as assumed in the model suggests that education concerning natural causes of both SP and its frightening hallucinations might be beneficial, though perhaps more so for more analytic individuals (see next section). These results suggest that it is not only important for clinicians to be aware of the implications of supernatural beliefs on SP distress but also sensitive to the likelihood that not everyone will accept or experience relief from naturalistic interpretations (cf. Hufford, 2005).
Historically, having unusual personal experiences, such as hearing voices, seeing things, and being assaulted and tormented by agents undetectable by others, has often been interpreted as a spiritual affliction. In many cultures and subcultures today, it still is (e.g., Adler, 2011; Hufford, 2005). Even in our sample, we see that such beliefs, though generally not strong, were fairly prevalent even in a presumably relatively sophisticated web sample. The inherent distress caused by these unpleasant experiences might therefore be increased by potential nocebo effects caused by the belief that these experiences also signify some spiritual danger or flaw in the victim’s character or failure of responsibility (Adler, 1991, 1994, 2011). Yet, there have been, throughout history and across cultures, remedies for such spiritual afflictions. Medieval priests in Europe, for example, had their “treatments” consisting of exorcism via prayer and liturgy, invocation of saints, communion, amulets, and flagellation. These may well have provided some relief via counteracting placebo effects of compelling ritual (Adler, 2011). With the development of the Enlightenment, our understanding of the experiences gradually changed—but only in some respects and not equally for all members, including intellectuals, of Western societies (Israel, 2010). Nonetheless, for many if not most, the major sources of these strange experiences have largely, though not entirely, changed from spiritual to material, and the notion of a possessed or corrupted spirit has been at least partially transformed into one of a diseased or disturbed mind. One consequence of this is that the old remedies are largely ineffectual, noncredible, and, in any case, often practically unavailable (Adler, 2011). Not believing in supernatural causes of one’s afflictions, given the absence of remedies, does seem to provide some measure of relief from, or at least reduced severity of, postepisode distress.
Analytic cognitive style was significantly negatively correlated with both supernatural beliefs about SP and SP postepisode distress independent of the measure of cognitive ability. This is consistent with previous studies in which analytic cognitive style was, independent of cognitive ability, negatively correlated with more general measures of supernatural (religious and paranormal) beliefs (Gervais & Norenzayan, 2012; Pennycook et al., 2012; Shenhav et al., 2012).  Obviously, the frightening nature of SP must often motivate the experient to seek some sort of explanation.
It has long been known that individual differences in cognitive styles may influence the kinds of information sought and their persuasive impact when found (e.g., Cacioppo & Petty, 1982; Cacioppo, Petty, & Morris, 1983). The characteristics of analytic individuals suggest that they may be more likely than the more intuitive individuals to seek out, and possibly be convinced by, naturalistic explanations and potentially be dissuaded from the supernatural interpretations, however strongly suggested by the experiences themselves. In addition, because there is no shortage of occult accounts of SP in both print and electronic media, more intuitive individuals may be drawn to, and their beliefs reinforced by, alternative explanations. Thus, it is entirely possible that self-selective exposure to media accounts of SP might exaggerate cognitive biases.
An unexpected result was the significant direct effect of analytic style on postepisode distress, rather than being completely mediated by beliefs. This required an elaboration of our model and requires an extension of our understanding of the effects of an analytic style on distress. If analyticity is, in part, the ability of executive control processes to inhibit or manage more automatic processes, this ability appeared, in the present study, to extend to the activation/inhibition of emotional reactivity. That is, a plausible explanation for the direct effect of analyticity on postepisode distress is that analytic people are more likely to override their postepisode feelings of distress. It is important to note that the argument being made here is not that more analytic individuals are less likely to experience fear. Fear during SP episodes had an effective zero correlation with analyticity. More analytic individuals were just as frightened during their SP experiences as less analytic individuals. The significantly greater negative correlation with postepisode distress revealed, however, that fear during episodes was less likely to be translated into postepisode distress. We have previously argued that analytic individuals may have the same intuitions as less analytic individuals but have more effective control processes for managing those intuitions (Pennycook et al., 2012). Similarly, in the SP context at least, analytic individuals appear to be no less emotionally reactive during episodes but are more likely to manage their emotional state following the episode. Such a possibility rather considerably extends the relevance of analytic cognitive style well beyond the current decision-making and belief-formation contexts (cf. Epstein et al., 1996).
Limitations and special considerations
Web samples, like all samples, have both advantages and limitations. Major advantages include the ability to acquire large samples that are much more diverse in age, nationality, and background than the (often student) samples used in a substantial proportion of SP studies. Yet, Web samples are self-selected rather than random. Such samples likely include participants who may be more knowledgeable than average, both generally and with regard to the specific topics under study. In addition, the present sample likely included individuals experiencing comparatively more frequent, intense, and elaborate hallucinatory experiences and who are likely to have experienced more distress than nonparticipants. Nonetheless, as we have noted previously, our Web samples have replicated our major findings, with a conventional student-based sample using traditional paper-and-pencil surveys, in an online Web sample (Cheyne & Girard, 2007). Web participants typically express considerable interest in, commitment to, and appreciation for our research, with over two thirds of participants in our general survey providing e-mail addresses so that they may be contacted for further research participation. We also suspect that because the research is perceived to be of a scientific nature, we may underestimate the incidence and intensity of paranormal beliefs about SP in the larger population.
SP experiences, beliefs, and distress were all assessed via self-report. Ultimately, all of these are inherently subjective, including SP experiences. The experiences, complaints, and distress considered here are ultimately “psychological,” whatever and however varied the underlying neurophysiological conditions might be. Moreover, although the surveys were rather detailed—especially the SP experiences, with multiple text boxes for participants to add comments and qualifications, which have helped us over the years develop more detailed questions and hypotheses—questionnaires do not, of course, provide the background of a full clinical workup. The purpose of the present study was to provide a broad-brush, multivariate picture of a substantial subset of potential sources of distress to contextualize a very particular sleep-related complaint.
Main clinical implications
A further advantage of online surveys is that text boxes can be provided for participants to provide qualitative information not considered by investigators; they also allow participants to contact investigators via e-mail, providing a two-way communication between investigator and participants. A frequent lament reported by participants in our studies over the years has been the lack of understanding and sometimes outright dismissiveness of the terror of the experiences and the distress they cause some individuals. Of particular clinical importance, perhaps, is the overdetermined nature of SP postepisode distress. Note in the model that episode characteristics, analytic cognitive style, supernatural beliefs, and general psychological distress were all orthogonal contributors to distress. This suggests a multidimensional approach to the evaluation of the sources of distress for any given individual. Does, for example, the patient’s distress reflect the severity of the SP episodes? If so, this suggests that the problem is indeed a sleep disorder–related phenomenon. Intense fear and vivid imagery during episodes appear to be relatively independent of the other variables contributing to distress, and so a focus on the sleep disorder seems in order. Hence, assessment of experiences during episodes is as critical as the assessment of postepisode distress. Yet, SP distress may be, in some cases, a symptom of problems beyond SP itself, reflecting a potentially diverse set of problems, including a more generalized psychological distress combined with beliefs that might catastophize a wide range of frightening anomalous experiences.
Although SP experiences do not strictly constitute a culture-bound syndrome (Cheyne, 2010), the present results do suggest that clinicians need to be sensitive to the importance of cultural beliefs, particularly with respect to the consequences of those beliefs about distressing experiences that have bizarre components that inherently suggest supernatural phenomena. Moreover, in dealing with such beliefs in the course of therapy, it is important to be aware of different cognitive styles, some of which may render the patient refractory to medical-scientific explanations.
Low levels of postepisode distress, such as rumination, are common, and levels of endorsement drop off fairly steeply as apparent severity of distress increases. The percentage of people reporting some effects on functioning the next day does suggest, however, that SP makes a significant contribution to the billions of dollars, worldwide, in costs associated with accidents, illnesses, and lost productivity associated with sleep disturbances. The present results indicate that postepisode distress is the result of a combination of the severity of the episodes, general psychological distress, cognitive style, and beliefs about the causes of SP. Although the nature of SP episodes clearly triggers postepisode distress, researchers, practitioners, and people who have SP episodes should all be sensitive to the possibility that beliefs and cognitive factors influencing those beliefs have substantial impact on subsequent distress. In summary, the present study provides a reliable scale to assess postepisode distress and initial evidence of the conceptual validity and utility of the scale. More generally, the present findings provide evidence of the contributions of both situational and individual difference factors to extremely stressful personal experiences.
Sleep Paralysis Supernatural Beliefs Scale
Here we ask for your opinion about some possible explanations of sleep paralysis experiences. Below are explanations people often give for such experiences. Please indicate how likely or unlikely each of the suggested explanations for sleep paralysis seems to you.
I consider it:
1 = highly unlikely 2 = unlikely 3 = Uncertain 4 = likely 5 = highly likely
that Sleep Paralysis can be explained as:
View this table:
In this window In a new window