Extinction learning of conditioned fear in adult Attention-Deficit and Hyperactivity Disorder (ADHD)

Introduction: Learning based on instruction can lead to similar fear adaptive responses alike learning by exposure to an aversive event itself [Phelps 2001]. Hence, learning by instruction prevents us from physical harm and from risk-taking behaviour. In ADHD risk-taking behaviour, like car speeding or substance abuse is well-known [Barkley 1996; Biederman 1998]. However, so far most of these symptoms have been explained in terms of insufficient impulse control. Yet, recent finding suggest changes in areas relevant for emotion processing. Plessen et al. [2006] found abnormal surface morphology of the amygdala's basolateral complex, an area important for fear conditioning [LeDoux 1990; Maren 2005], in children. This study examined ADHD patients' fear learning circuitry in an instructed fear (IF) paradigm and tested for differences in functional and electrophysiological correlates during fear extinction.

Methods: In the IF paradigm subjects were first familiarised with an aversive electrical stimulus (AS) and adjusted the AS intensity in an dial-up procedure to an unpleasant but not painful level. Subsequently, subjects were instructed that they can experience an AS when viewing a square of one of two different colours. After instruction subjects viewed blue and yellow squares in a randomised sequence, while their skin conductance and BOLD signal was recorded. After the first session subjects were reinstated and underwent a further FE sequence. So far 11 ADHD patients and 14 healthy controls have been included.

Results: Interaction analysis (group x condition x time) of brain regions previously shown to be associated with extinction (ROIs) did not show the expected decline in amygdala activation in ADHD patients, but rather displayed an increasing response across time. Patients also activated the amygdala significantly higher for the early safe condition. In late extinction this activation dropped to a similar level as in controls. While controls showed a strong increase in rostrolateral orbitofrontal Cortex (rlOFC) in late compared to early extinction of threat, this effect was reversed in ADHD. Interestingly, patients tended to show an anterior insula (AIC) and dorsal anterior cingulate (dACC) hypoactivity.

Discussion: Differences in amygdalar, ACC and OFC activation go along with earlier volumetric and functional findings [Bush 1999; Hesslinger 2002; Plessen 2006; Seidman 2006] indicating deficits in brain areas important for emotion processing and regulation. Amygdalar hyperactivity in regard to the threat condition could be related to a diminished regulatory influence of the rlOFC. AIC and dACC hyporesponsiveness might indicate a diminished cognitive processing of fearful stimuli. In conclusion, these results suggest difficulties in assessing harmful stimuli, and a deficit in emotion regulation.