Attention-Deficit/Hyperactivity Disorder (ADHD), a behavioral disorder affecting 3-5% of school-age children, is predominantly characterized by developmentally inappropriate patterns of inattention, hyperactivity, and impulsivity [1]. Results from studies of molecular genetics, pharmacological effects of stimulant drugs, and neuroimaging suggest that dopamine dysfunction is an important factor in ADHD etiology [2, 3]. ADHD is a highly heritable disorder with an estimated heritability of ~0.76 [4]. Behavioral variance in heritability research is in a population often categorized into a genetic factor and an environmental factor. The gene-environment interaction (G × E), which is difficult to estimate, is usually incorporated into the genetic factor which inflates the estimate of purely genetic influences on ADHD prevalence and masks gene-environment interactions [5, 6]. Thus, environmental factors may alone contribute to 20% of ADHD prevalence, and additionally may interact with a genetic vulnerability and produce ADHD [7]. Environmental risk factors identified thus far are cigarette smoking and alcohol use during pregnancy, and low birth weight [7–9]. Another environmental factor that either alone or in concert with a genetic vulnerability may contribute to ADHD prevalence is exposure during early development to various environmental toxic substances like lead, mercury or polychlorinated biphenyls (PCBs) [10–15].

PCBs are a group of chemical industrial compounds consisting of 209 congeners which have been used as oils and coolants in electrical equipment as well as in building material [16]. Of the estimated 1.2 million tons total production worldwide, 30% was discharged to the environment [17, 18]. PCBs were banned around 1980 due to growing evidence that these compounds have adverse effects. They are chemically stable and resistant to degradation, and worldwide many areas continue to be heavily polluted. PCBs are stored in body-fat, accumulate in the food chain, and humans may be exposed prenatally both via trans-placental transfer and through breast milk during infancy. Adolescents and adults are mainly exposed through consumption of contaminated food, of which fish and seafood constitute the most important sources of PCB [16, 17, 19]. Studies show that even low-level exposure to PCBs during development has adverse effects on neurobiological, cognitive, and behavioral functioning [16, 20, 21]. Exposure may lead to impulsivity, reduced attention and concentration, poorer working memory and lower IQ scores [13, 22–28], and give rise to behavioral impairments that share significant similarities with ADHD [15]. Studies have shown an association between PCB-levels in the umbilical cord blood and ADHD-behaviors as measured by Conners’ rating scale for teachers [13]. Moreover, an association between levels of cord serum PCBs and a higher omission rate on the Continuous Performance Test (CPT) and slower processing speed on WISC-III was found in boys, whereas for girls, the association was in the opposite direction for the CPT and zero for the WISC-III [29]. These latter findings suggest that exposure to PCBs may interact with gender to produce different behavioral and cognitive outcomes in males and females. Consistent with research on humans, studies show that PCB exposure affects learning and memory, activity level, and cognitive functions also in animals (for reviews, see [16, 19]).

The dynamic developmental theory of ADHD [30] proposes that fundamental learning processes are altered in ADHD, consistent with suggestions of a reinforcement deficit in ADHD [31–41]. The dopamine system plays a major role in reinforcement and extinction of behavior which are the fundamental mechanisms of behavioral selection [42, 43]. Behavior is selected by its consequences: Reinforcers increase the probability of behavior that produced them, and inadequate behavior is eliminated through the extinction process. From the selection and combination of simple behavioral units into longer behavioral sequences, habits and skills are built, and these behavioral sequences come under the control of environmental stimuli (stimulus control) as behavior has different consequences across environmental settings. According to the dynamic developmental theory of ADHD, dopamine dysfunction reduces the temporal window for associating antecedent stimuli and behavior with the behavior’s consequences (the three-term contingency), and weakens the elimination of inadequate behavior (extinction). These changes in fundamental learning processes are proposed to produce the behavioral symptoms of inattention (lack of stimulus-control), hyperactivity, and impulsivity (excess of responses with short inter-response times) observed in ADHD [44]. Additionally, the theory suggests that the sequencing of behavioral units into orderly, predictable chains of behavior is deficient in ADHD, producing short chains of behavior with overall low predictability. The predictability of behavioral chains in ADHD has been investigated using autocorrelations, where serial correlations between instances of behavior are calculated across the entire data set (e.g. the correlation between response n and response n + 1, between response n and response n + 2, …). These studies have found lower autocorrelations in the behavior of children with ADHD [45, 46].

The present study examined the hypothesis that a relation exists between exposure to PCB 153 (2,2′,4,4′,5,5′-Hexachlorobiphenyl), a PCB congener found in human breast milk, and ADHD-like behavior in the spontaneously hypertensive rats (SHR/NCrl), an animal model of ADHD, and in Wistar Kyoto controls [47–49]. The SHR/NCrl represents a multifactorial genetic rat strain, and is with Wistar Kyoto (WKY/NHsd) controls a well validated animal model of ADHD [50, 51]. Previous findings suggest that the behavioral changes observed in SHR/NCrl are linked to imbalanced dopamine and noradrenalin transmitter systems, both of which are affected by PCB exposure [48, 52–56].

It was therefore hypothesized that the SHR/NCrl would be sensitive to PCB exposure to a greater extent than the WKY/NHsd controls. Findings suggest that PCB exposure activates compensatory brain mechanism [57, 58], and a pre-existing transmitter system imbalance in SHR/NCrl may reduce the effectiveness of these compensatory brain mechanisms and lead to more behavioral changes following PCB exposure than in normal controls. We also wanted to explore if exposure would produce ADHD-like behaviors in WKY/NHsd control animals. In the present study, female and male animals were included to test if exposure effects were similar across sex, and short-term and long-term behavioral changes following exposure were investigated. Finally, three PCB doses were included to examine the dose–response relations in male and female SHR/NCrl and WKY/NHsd.

The rats were exposed to PCB 153 three times between postnatal day (PND) 8 and 20, which is a period of major brain development. In rats, the growth spurt of the brain peaks around PND 7 and is largely completed by PND 20. The corresponding time interval in humans is from birth to age 20, and must be considered when interpreting findings from translational research [59]. To accurately control dosing, the animals were in the present study exposed post-natally using a stomach tube, and effects of three doses (1, 3 or 6 mg/kg) representing low, medium, and high dose exposure were tested. Effects of PCB 153 exposure were subsequently assessed by reinforcer-controlled lever pressing representing operationalizations of deficient sustained attention (stimulus control), hyperactivity (lever presses with IRTs > 0.67 s), and impulsivity (responses with IRTs < 0.67 s). This behavioral procedure has been used in a number of previous studies of SHR/NCrl and WKY/NHsd including studies of drug-effects and PCB exposure [47, 48, 51, 60, 61]. As a supplement to lever-pressing, long-term effects of PCB exposure were in selected sessions assessed by video-recordings of the animal’s behavior during operant testing. Whenever the animal moved, the pixels changed from one video-frame to the next. These pixel-changes were used to calculate how amount of movement during testing. Additionally, we examined possible PCB exposure effects on the predictability of the sequential movement pattern by autocorrelating the animal’s positions in the operant chamber across the duration of the session. The behavioral testing of PCB exposure effects was completed before PND 70 to avoid the confounding factor of hypertension that SHR/NCrl develop from 10 to 12 weeks of age [62].

The main aim of the study was to test effects of PCB 153 exposure on ADHD-like behaviors in the animal model of ADHD, the Spontaneously hypertensive rats (SHR/NCrl), and in Wistar Kyoto (WKY/NHsd) controls. Also, sex-specific effects and dose–response relations were explored. For this purpose, SHR/NCrl and WKY/NHsd rats, males and females, were exposed three times between PND 8 and 20 to PCB 153 at doses of 1, 3, or 6 mg/kg bodyweight or to corn oil only and then behaviorally tested from PND 37 to PND 64 for changes in stimulus control (sustained attention), number of lever-presses with IRTs > 0.67 s (activity), and number of lever-presses with IRTs < 0.67 s (impulsivity). Additionally, the animals were video-recorded during the operant task at PND 54 and 63, and amount of movement and the predictability of the movement pattern (autocorrelation of the animal’s position) in the operant chamber were analyzed.

The major short-term PCB 153 exposure effects include observations of: 1) No short-term effects of PCB exposure on stimulus control (attention) in SHR/NCrl. In WKY/NHsd, exposure to 1 mg/kg was associated with a short-term decrease in stimulus control that was more pronounced in males than in females. Still, a statistically significant difference was only found between the 1 mg/kg and 3 mg/kg WKY/NHsd exposure groups (Figure 1). 2) In SHR/NCrl, PCB exposure was associated with a short-term decrease in responses with long and short IRTs (activity and impulsivity, respectively) in animals exposed to 1 mg/kg. However, significant effects were only found for comparison of activity in the 1 mg/kg and 6 mg/kg SHR/NCrl exposure groups, and for comparisons of impulsivity in the 1 mg/kg with the 3 mg/kg or 6 mg/kg SHR/NCrl exposure groups. In contrast, activity was increased in WKY/NHsd exposed to 1 mg/kg and was significantly different from WKY/NHsd controls and animals exposed to 3 mg/kg, whereas no effects were found on impulsivity (Figures 2 and 3).

The major long-term PCB 153 exposure effects include observations of: 1) A long-term increase in stimulus control in male SHR/NCrl exposed to 1 mg/kg, whereas this effect was transient in SHR/NCrl females. Additionally, decreased stimulus control was observed in SHR/NCrl females exposed to 3 mg/kg and 6 mg/kg. However, significant differences were only found for comparisons of the 1 mg/kg SHR/NCrl exposure group with the 3 mg/kg or 6 mg/kg SHR/NCrl exposure groups. No long-term exposure effects on stimulus control were observed in WKY/NHsd (Figure 1). 2) A long-term decrease in activity in SHR/NCrl exposed to 1 mg/kg relative the other SHR/NCrl exposure conditions, and decreased impulsivity in SHR/NCrl exposed to 1 mg/kg compared to SHR/NCrl exposed to 3 mg/kg or 6 mg/kg. No long-term effects on these measures were observed in WKY/NHsd (Figures 2 and 3). 3) A long-term decrease in video-recorded movement in SHR/NCrl exposed to 1 mg/kg relative to SHR/NCrl exposed to 3 mg/kg or 6 mg/kg. However, exposure effects were mainly observed in SHR/NCrl females: Unexposed females moved significantly less than females exposed to 3 mg/kg, and females exposed to 1 mg/kg moved significantly less than females exposed to 3 mg/kg or 6 mg/kg. In contrast, no exposure effects on amount of movement were observed in WKY/NHs (Figure 4). 4) The autocorrelations of the position in the chamber tended to be higher in SHR/NCrl exposed to 1 mg/kg compared to unexposed controls. Also, the autocorrelations in SHR/NCrl exposed to 6 mg/kg and in SHR/NCrl females exposed to 3 mg/kg tended to be lower than in unexposed controls. The autocorrelations in unexposed WKY/NHsd males tended to be lower than in males exposed to 3 mg/kg and higher than in males exposed to 1 mg/kg, whereas no effects were observed in WKY/NHsd females. The statistical analyses showed significantly higher autocorrelations in SHR/NCrl exposed to 1 mg/kg compared to the 6 mg/kg exposure group. In contrast, the autocorrelations were significantly lower in WKY/NHsd exposed to 1 mg/kg compared to WKY/NHsd exposed to 3 mg/kg. Also, statistically significantly higher autocorrelations were found in males exposed to 3 mg/kg compared to males exposed to 6 mg/kg, whereas no significant differences were found in females (Figure 5).

The present findings further indicate that video-derived measures may supplement observations of lever pressing for the general purposes of testing behavior-altering effects of PCB 153 exposure (Figures 4 and 5). A previous study using an identical experimental procedure found a dissociation between video-recorded movement and rate of lever-pressing, suggesting that these measures tap into different behaviors [70]. A similar effect was found in the present study. PCB 153 exposure was associated with changes in autocorrelations of the animal’s position in the operant chamber although no changes were observed for lever-pressing, especially in the males. This dissociation can be seen by comparing Figures 1, 2, 3 (video-analyzed sessions are marked in red on the x-axes) with Figure 5 for doses 3 mg/kg and 6 mg/kg, and support the suggestion that two different aspects of behavior are measured. Generally, the predictability of the sequential movement pattern as analyzed by autocorrelations of the animal’s position in the test chamber shows that predictability is lower in SHR/NCrl, consistent with the dynamic developmental theory of ADHD and with findings in children with ADHD [30, 45, 46]. This finding is also consistent with studies of SHR/NCrl using other measures of reduced behavioral predictability (e.g. intra-individual variability, entropy) [71, 72].

Taken together, our data show that exposure to PCB 153 produced a complex pattern of effects that depended on strain, sex, dose, time of testing, and the behavioral measure used. Generally, exposure produced long-lasting and larger behavioral changes in SHR/NCrl compared to WKY/NHsd controls, with mainly small and transient behavioral changes observed in the latter. Compared to unexposed controls and for many of the behavioral measures, the observed exposure effects of one dose were in the opposite direction to that observed following a higher dose, especially for SHR/NCrl. For several of the post hoc analyses, statistically significant differences were found only between two or more doses, whereas few significant effects were found between one individual dose and the unexposed control condition. This lack of a statistically significant difference between the exposed and unexposed groups complicates the interpretation of the results. Still, the data show a consistent pattern of effects across all behavioral measures that seems an unlikely chance occurrence. Overall, the data indicate a non-linear dose–response relationship with opposite effects of the doses tested, and where the effects reached the level of significance only when these opposite effects were compared.

The diversity of experimental procedures used in a number of published PCB studies precludes a direct comparison with our findings. Possibly, most relevant to the present results due to similarities in experimental procedure are the studies by Holene et al. which observed a long-lasting decrease in stimulus control and an increase in activity and responses with short IRTs in male, but not female, offspring of DA/OLA/HSD females mated with Lewis rats and gavage-fed 5 mg/kg PCB 153 every second day from PND 3 to 13 [73, 74]. Comparable effects were found by Berger et al. in adult male Sprague–Dawley rats exposed to low doses of the PCB mixture Aroclor 1248 (30 g daily portions of rodent diet supplemented with 1 ml corn oil containing 0.5 μg/g Aroclor 1248) in the diet during puberty [75]. In the present study, similar behavioral patterns following PCB exposure were observed in female SHR/NCrl exposed to 3 and 6 mg/kg, and partly in male WKY/NHsd following exposure to 1 mg/kg (Figures 1, 2, 3). The combined result patterns indicate that a number of variables interact with PCB exposure, and that the neurotoxic effects on the neural system may differ across strain, sex and dosing regimen.

PCB 153 exposure was associated with pronounced and long-lasting behavioral changes in SHR/NCrl, suggesting greater behavioral sensitivity to PCB exposure in these animals as compared to WKY/NHsd controls. PCB 153 exposure did not unequivocally aggravate ADHD-like behaviors in SHR/NCrl but depended on dose where 1 mg/kg tended to reduce ADHD-like behaviors and produce opposite behavioral effects compared to 3 mg/kg and 6 mg/kg, especially in the SHR/NCrl females. For the WKY/NHsd controls and for the three doses tested, PCB 153 exposure did not produce the full range of ADHD-like behaviors which were observed in unexposed SHR/NCrl, but produced a few specific behavioral changes in males only.

To the extent the present findings can be generalized to human ADHD, exposure effects of PCB 153 on ADHD behavior depend on amount of exposure: Low doses moderate and high doses aggravate ADHD symptoms in genetically vulnerable individuals. In normal controls, exposure does not seem to constitute an environmental risk factor for developing ADHD, but can produce specific behavioral changes.

At a general level and consistent with previous findings, the present findings suggest that behavioral changes can be related to dose in an non-linear fashion, and that exposure to PCB 153 interacts with several variables, including strain, sex, dose, and time of testing.