Information Processing Capacity in Psychopathy: Effects of Anomalous Attention

Abstract

Hamilton and colleagues (2015) recently proposed that an integrative deficit in psychopathy restricts simultaneous processing, thereby leaving fewer resources available for information encoding, narrowing the scope of attention, and undermining associative processing. The current study evaluated this parallel processing deficit proposal using the Simultaneous – Sequential paradigm. This investigation marks the first a priori test of the Hamilton et al.’s theoretical framework. We predicted that psychopathy would be associated with inferior performance (as indexed by lower accuracy and longer response time) on trials requiring simultaneous processing of visual information relative to trials necessitating sequential processing. Results were consistent with these predictions, supporting the proposal that psychopathy is characterized by a reduced capacity to process multicomponent perceptual information concurrently. We discuss the potential implications of impaired simultaneous processing for the conceptualization of the psychopathic deficit.

Information processing involves a cascade of neurobiological events that transforms sensory input into neural representations and subsequent behavior. Failure at any step in the processing sequence would therefore influence the content and quality of perceived information and thus the quality of one’s behavioral response (Goldstein, 2010). Psychopathy is a clinical syndrome typified by socially and legally significant antisocial behavior. High levels of psychopathic traits are predictive of engagement in criminal behavior, violent offending, and high rates of recidivism (Hare & McPherson, 1984). Accordingly, specifying the information processing abnormalities of psychopathic individuals may provide crucial insight into their behavioral dysfunction.

Even before its formalization as a diagnostic entity, the psychopathy construct has long been recognized in historical literature and lore (Buzina, 2012). A commonality across both early and modern conceptions is the central role of moral degeneracy and impulsive, aggressive behavior. Deficient affective experience is traditionally considered the hallmark of the psychopathic syndrome (e.g., Blair, Mitchell, & Blair, 2005; Lykken, 1995).

Despite the centrality of affective dysfunction in psychopathy, research suggests that the construct is associated with broader cognitive deficiencies, suggesting that information processing abnormalities may subsume emotion deficits. These findings have spawned numerous theories positing diverse information processing deficits in psychopathy (e.g., Baskin-Sommers, Curtin, & Newman, 2011; Kosson, 1996; Patterson & Newman, 1993). Nevertheless, elucidation of the dysfunction that accurately predicts the full range of information processing strengths and weaknesses in psychopathy has proven elusive (see Hamilton, Hiatt Racer, & Newman, 2015 for a review of prevailing emotion and cognitive models and their shortcomings).

Hamilton and colleagues (2015) recently proposed the Impaired Integration (II) theory of psychopathy, a framework that makes more explicit the neurological underpinnings and dysfunctional neurocognitive processes in psychopathy. The II theory leads to a variety of novel, testable predictions that concurrently build on and deviate from those from previous etiological theories. The II theory proposes that the fundamental deficit in psychopathy involves difficulty rapidly integrating multicomponent information. As a result, it is more difficult for psychopathic individuals to encode multiple stimuli or multiple aspects of stimuli simultaneously, bind spatially and temporally discontiguous information, and interpret perceptual input based on associative connections in long-term memory. Moreover, stemming from these limitations, psychopathic individuals have reduced attentional capacity for active encoding, maintenance of relevant stimuli, and integrating events and their contexts of occurrence with existing schemas. Thus, a central tenet of the II framework is that simultaneous processing capacity will be relatively impaired in psychopathy; accordingly, it predicts that psychopathic individuals will benefit from sequentially-presented multicomponent information.

One paradigm that is especially pertinent for examining differences in information processing capacity is the Simultaneous-Sequential paradigm. In this paradigm, participants are instructed to make judgments about briefly presented stimuli that are either presented simultaneously (i.e., all at once) or sequentially (i.e., subsets are presented one-at-a-time). Accordingly, the critical manipulation in this task is the number of simultaneously presented stimuli. Processing capacity limits are task-dependent and reflect limitations of perceptual encoding and/or sematic categorization (Scharff, Palmer, & Moore, 2011a). These limits influence whether humans process sensory input simultaneously or sequentially on a given task. Typically, there is a selective advantage for sequential processing on tasks with low target-distracter discriminability (i.e., greater difficulty; Scharff, Palmer, & Moore, 2011b). The Simultaneous-Sequential paradigm was originally developed to study the rate of information processing in visual perception (Eriksen & Spencer, 1969) and since its conception has been used primarily to answer questions regarding visual capacity limitations (see Scharff et al., 2011b).

The current study sought to examine psychopathy-related differences in processing capacity for simple visual stimuli; more specifically, it aimed to elucidate the extent to which psychopathy is preferentially associated with sequential relative to simultaneous processing. The use of stimuli with intermediate target-distracter discriminability enabled us to accommodate the abilities of our sample while maximizing task sensitivity to individual differences in processing style. Based on the assumption that psychopathy is characterized by information processing restrictions that foster the sequential processing of information, we hypothesize that visual processing accuracy and response time in psychopathy will be moderated by information presentation style. Specifically, we predict that psychopathic individuals will have lower accuracy and slower response times on simultaneous trials and higher accuracy and faster response times on sequential trials.

Methods

Participants

Participants consisted of 87 Caucasian male inmates ages 20 to 55 (M = 31.03, SD = 8.22) from a medium-security prison in central Wisconsin. To be included in the study, participants had to be between 18 and 55 years old, free of a history of psychosis or bipolar disorder, not currently taking psychotropic medication, and have an IQ score of 70 or greater¹. All participants provided written informed consent according to procedures approved by the University of Wisconsin – Madison Human Subjects Committee. On the first day of the study, interviewers conducted semi-structured life history interviews with the inmates. This interview included questions on childhood, education, and occupational, interpersonal, and legal histories. Interviewers then reviewed institutional files to corroborate information provided during the interviews. The combination of interview and file information was used to rate psychopathy using to Hare’s (2003) Psychopathy Checklist-Revised (PCLR). One participant was excluded from analyses due to low accuracy on the experimental task (less than 60% correct across all trial types).

Psychopathy Checklist-Revised (PCL-R; Hare, 1991)

The PCL-R consists of 20 items that are rated according to the degree to which a characteristic is present (significantly = 2, moderately = 1, not at all = 0). In the present sample, scores on this measure ranged from 10 to 34, with a mean of 22.51 (SD = 6.03). Interrater reliability (intraclass correlation) for PCL-R total score, based on six dual ratings, was .96.

Materials

Stimuli

Stimuli were 360 sets of four letters written in white font embedded in Gaussian noise patches presented on a gray background. Each set of letters consisted of three distractor stimuli (A, C, F, L, M, N, P, Q, T, V, W, X, or Z) and one target stimulus (B, G, H, R, or Y) that varied by block.

Experimental Task

The simultaneous-sequential paradigm consists of a visual search task in which four stimuli, one of which is the predefined target, are presented either simultaneously or sequentially. In the current task, participants were presented with a target letter at the beginning of each block. Following the target display, participants saw a fixation cross that was followed by the presentation of four letters presented at once or divided equally in two separate frames. Participants were instructed to indicate the location of the target letter as quickly and accurately as possible.

The task consisted of 75 practice trials and 720 experimental trials. The first 45 practice trials consisted of 15 simultaneous, 15 sequential, and 15 repeated trials presented in isolated blocks so that participants could become familiar with each condition. On simultaneous trials, all letter stimuli were displayed in one 100 ms frame. On sequential trials, the stimuli were divided equally between two 100 ms frames that were separated by a 1 s fixation frame. On repeated trials, all four stimuli were presented simultaneously twice in two 100 ms frames that were separated by a 1 s fixation frame (see Figure 1). The repeated trials were included based on Scharff et al.’s (2011b) extended paradigm to address capacity processing models. Since this condition is not related to the present hypotheses, results pertaining to this condition are discussed in supplemental analyses.² Once all stimuli were presented, participants saw a cue screen that read ‘Location?” indicated the target location via a button press. This screen appeared until participants responded or 1 s had elapsed. For the remaining 30 practice trials, all trial types were intermixed. To ensure a lack of practice effects on experimental trials, a unique target letter (U) was used in practice sessions. The 720 experimental trials were divided into ten blocks (72 trials each) of randomized trial types; each block was separated by a 10-s break. The duration of the practice and experimental trials was approximately 40 minutes.

Figure 1.

Trial structure and experimental conditions. On each trial, participants saw a series of four letters, one of which was a predefined target. Letter presentation varied by condition. (a) Simultaneous trials: all four stimuli were displayed in one 100 ms frame. (b) Sequential trials: stimuli were divided equally between two 100 ms frames separated by a 1 s fixation. (c) Repeated trials: all four stimuli were displayed in two 100 ms frames separated by a 1 s fixation. Participants were told to indicate the location of the target letter by pressing one of four designated buttons on a keyboard. The trial ended after the participant responded or after 1 s. Each trial was followed by a 250 ms inter-trial interval.

Data Analysis

The goal of the current study was to examine psychopathy-related differences in perceptual processing capacity. To test for psychopathy-related differences in performance we used two general linear models (GLMs) with trial type (sequential vs. simultaneous) and block (early vs. late) as repeated measures, z-scored PCL-R total score³ as a continuously distributed between-subject factor, and accuracy and response time as respective dependent variables. To control for differences in overall performance, we included z-scored averaged simultaneous-sequential accuracy and response time as continuously distributed covariates in the respective analyses.^{4, 5}

To characterize significant interaction effects, we used point estimates generated from the GLMs to estimate the conditional effect (i.e., simple slope) of trial type on mean accuracy for individuals high and low in psychopathy (1.25 SD above and below the sample mean PCL-R total score, respectively⁶). We repeated these analyses using response time as the dependent variable. Partial eta squared values are included as measures of effect size.

Results

Covariate-adjusted mean accuracy by trial type and estimated mean values for individuals high and low in psychopathy are presented in Table 1. Table 2 provides the inter-correlations among the study variables.

Table 1.

Covariate-Adjusted Mean Accuracy and Response Time by Trial Type

	Condition mean		High psychopathy		Low psychopathy
	ACC	RT	ACC	RT	ACC	RT
SEQ	.86 (.01)	584.72 (5.10)	.88 (.02)	567.05 (8.21)	.83 (.02)	602.39 (8.21)
SIM	.81 (.01)	792.23 (5.10)	.78 (.02)	809.90 (8.21)	.83 (.02)	774.55 (8.21)

Note: Psychopathy scores were obtained using the Psychopathy Checklist-Revised (PCL-R; Hare, 1991). Since we analyzed psychopathy continuously rather than using an extreme-group design, the values presented are point estimates (i.e., estimated using regression analyses) for low and high psychopathy points (1.25 SD below and above the sample mean PCL-R total score, respectively) on the distribution. Numbers in parentheses are standard errors. SEQ = performance on sequential trials; SIM = performance on simultaneous trials.

Table 2.

Bivariate Correlations (r values) between PCL-R Total Score and Performance Variables by Trial Type

	PCL-R	SEQ ACC	SIM ACC	SEQ RT	SIM RT	Diff ACC	Diff RT
PCL-R	-
SEQ ACC	.16	-
SIM ACC	−.12	.27*	-
SEQ RT	−.04	.17	−.38**	-
SIM RT	.18	.31**	−.48**	.71**	-
Diff ACC	.22*	.57**	−.63**	.45**	.66**	-
Diff RT	−.29**	−.20	.14	.38**	−.39**	.28**	-

Note. The difference scores included in the table represent the difference in accuracy and response time on sequential trials compared to simultaneous trials. These variables allow for the characterization of the magnitude of the association between study variables and the differential performance on simultaneous relative to sequential trials. PCL-R = psychopathy total score; SEQ ACC = accuracy on sequential trials; SIM ACC = accuracy on simultaneous trials; SEQ RT = response time on sequential trials; SIM RT = response time on simultaneous trials; Diff ACC = represents the difference in sequential trial accuracy relative to simultaneous trial accuracy, computed by subtracting accuracy on simultaneous trials from accuracy on sequential trials; Diff RT = represents the difference in response time to sequential trials relative to that on simultaneous trials, computed by subtracting response time on simultaneous trials from response time on sequential trials;

^*p < .05;

^**p < .01

Psychopathy - Accuracy Analyses

There was a significant main effect of trial type (F(1, 84) = 6.62, p = .01, η_p² = .07), such that accuracy on sequential trials was significantly higher than accuracy on simultaneous trials. Moreover, there was a significant block effect (F(1, 84) = 74.23, p < .01, η_p² = .47), with participant accuracy increasing over the course of the task. The main effect of PCL-R score on task accuracy was not significant (F(1, 85) = .16, p = .69, η_p² < .01). The main effect of trial type was moderated by psychopathy level, F(1, 84) = 4.35, p = .04, η_p² = .05. Follow-up analyses estimating the effect of trial type for individuals high and low in psychopathy indicated that high psychopathy scores were associated with significantly greater accuracy on sequential trials relative to simultaneous trials (F(1, 84) = 10.45, p < .01, η_p² = .11). In contrast, accuracy across trial types did not differ for individuals with low psychopathy scores, F(1, 84) < .01, p = .98, η_p² < .01 (see Figure 2). There was no interactive effect of block and PCL-R score on accuracy (F(1, 84) = 1.60, p = .21, η_p² = .02), suggesting that changes in accuracy across block was comparable across psychopathy level. The three-way interaction between psychopathy score, block, and trial type was also not significant, F(1, 84) = 2.12, p = .15, η_p² = .03.

Figure 2.

Accuracy by trial type for individuals high and low in psychopathy. Predicted values (point estimates) for task accuracy are depicted as a function of psychopathy level and trial type. High and low psychopathy levels are defined as the sample mean PCL-R score plus or minus 1.25 SD, respectively. Error bars represent standard error.

Psychopathy - Response Time Analyses

There was a significant main effect of trial type on response time (F(1, 84) 413.78, p < .01, η_p² = .83), such that responses on sequential trials were significantly faster than on simultaneous trials. There was also a significant block effect (F(1, 84) = 50.57, p < .01, η_p² = .38), with mean response time decreasing over the course of the task. There was no significant main effect of PCL-R score on response time (F(1, 85) = .65, p = .42, η_p² < .01). Consistent with hypotheses, the main effect of trial type was qualified by psychopathy, F(1, 84) = 7.54, p = .01, η_p² = .08. Subsequent analyses estimating the effect of trial type for individuals high and low in psychopathy demonstrated that although the effect of trial type on response time was significant among individuals with low psychopathy scores (F(1, 84) = 109.90, p < .01, η_p² = .57), this effect was larger among high psychopathy participants, F(1, 84) = 218.74, p < .01, η_p² = .72 (see Figure 3). Neither the block by psychopathy interaction (F(1, 84) = .05, p = .83, η_p² < .01) nor the three-way interaction between psychopathy score, block, and trial type was significant, F(1, 84) = 1.18, p = .28, η_p² = .01.

Figure 3.

Response time by trial type for individuals high and low in psychopathy. Predicted values (point estimates) for response time are depicted as a function of psychopathy level and trial type. High and low psychopathy levels are defined as the sample mean PCL-R score plus or minus 1.25 SD, respectively. Error bars represent standard error.

Discussion

The primary goal of the current study was to test the hypothesis that psychopathic individuals process multicomponent visual information in a serial versus parallel fashion. The II theory predicts that psychopathy is characterized by an information-processing deficit that constrains the amount of information that can be processed in parallel. As a result, psychopathic individuals are expected to be relatively impaired in processing simultaneously presented stimuli and to benefit preferentially from the serial presentation of multicomponent information. As predicted, psychopathy was associated with poorer performance on simultaneous trials relative to sequential trials.

The presented results are consistent with a growing body of evidence indicating that psychopathy is characterized by information processing anomalies that extend beyond emotion. According to the II perspective, the affective and disinhibitory symptoms seen in psychopathy are a downstream consequence of difficulty rapidly integrating components of multidimensional sensory stimuli. That is, inefficient processing undermines psychopathic individuals’ use of affective and inhibitory cues, especially under time limitations.

The II theory is a useful framework for understanding psychopathic abnormalities within the context of the Research Domain Criteria (RDoC) initiative at The National Institute of Mental Health. It makes predictions at the neurobiological, cognitive, and behavioral levels, with the current study focusing on the cognitive domain. Other investigations are specifying cognitive correlates in psychopathy by examining how attention influences perception and memory, in addition to testing predictions of neural modularity.

Overall, the current study offers explicit support for the theoretical premise of the II framework (i.e., relatively impaired simultaneous processing), and outlines a promising avenue for future research. Although further replication is needed to verify the reliability of the current findings, the proposal that psychopathy is characterized by a deficit in the ability to rapidly process simultaneously-presented information has potentially important implications for therapeutic interventions. Specifically, if future evidence associates this deficit with symptoms of psychopathy, treatments focusing on general information processing constraints may be effective in reducing psychopathic disinhibition.

Limitations

Before concluding, it is important to note that the current study was not designed to differentiate the II theory and attention bottleneck (AB) model, and the findings are consistent with both models. The AB model holds that psychopathy is characterized by an attentional bottleneck that impedes effective response modulation, precluding the processing of secondary information that is incongruent with a primary focus of attention (see Newman & Baskin-Sommers, 2011). In the current task, all stimuli are perceptually similar and related to an explicit, primary (i.e., top-down) focus of attention. Although the perspectives are overlapping, the II theory is more specific and direct in generating the a priori prediction addressed in this report.

Another potential concern is that the correlations between psychopathy and performance within conditions are small and generally non-significant (see Table 2). However, the sensitivity and validity of these measures are more readily apparent through examination of the differential effects trial type on varying psychopathy levels. The differences in accuracy and response time to sequential relative to simultaneous trials are captured through difference scores in which simultaneous trial performance is subtracted from sequential trial performance. For this reason, we have reported these difference scores in Table 2.

Another potential limitation relates to the specificity of the studied population. The present sample consisted exclusively of European American male prisoners. The homogeneity of the sample limits the generalizability of the presented findings. While prior research suggests that the presence of specific laboratory correlates is not consistent across sexes (Vitale, Maccoon, & Newman, 2011) or racial groups (see Smith & Lilienfeld, 2015), studying psychopathy in diverse populations is important for complete understanding of the construct. Thus, future research should explore perceptual processing capacity in psychopathic individuals of different sexes and from different racial backgrounds.

Conclusion

Overall, the current study provides novel support for the II theory. The finding that psychopathic individuals appear to benefit from the sequential presentation of information suggests that strategies for reducing psychopathic behavior should accommodate a serial processing style. Further research is needed to characterize the neurobiological correlates of impaired processing of simultaneously presented information. Neuroimaging studies will be useful in specifying the mechanism underlying this deficit. Moreover, future research should focus on the developmental trajectory of information processing deficits in psychopathy. Evidence of this deficit in children with psychopathic traits could enable early identification and remediation of potentially maladaptive information processing styles. Overall, research on the neurobiological underpinnings of cognitive abnormalities in psychopathy is necessary for developing a more precise understanding of and more effective treatments for psychopathic individuals.