Separate amygdala subregions signal surprise and predictiveness during associative fear learning in humans

Abstract

It has recently been suggested that learning signals in the amygdala might be best characterized by attentional theories of associative learning [such as Pearce-Hall (PH)] and more recent hybrid variants that combine Rescorla-Wagner and PH learning models. In these models, unsigned prediction errors (PEs) determine the associability of a cue, which is used in turn to control learning of outcome expectations dynamically and reflects a function of the reliability of prior outcome predictions. Here, we employed an aversive Pavlovian reversal-learning task to investigate computational signals derived from such a hybrid model. Unlike previous accounts, our paradigm allowed for the separate assessment of associability at the time of cue presentation and PEs at the time of outcome. We combined this approach with high-resolution functional magnetic resonance imaging to understand how different subregions of the human amygdala contribute to associative learning. Signal changes in the corticomedial amygdala and in the midbrain represented unsigned PEs at the time of outcome showing increased responses irrespective of whether a shock was unexpectedly administered or omitted. In contrast, activity in basolateral amygdala regions correlated negatively with associability at the time of cue presentation. Thus, whereas the corticomedial amygdala and the midbrain reflected immediate surprise, the basolateral amygdala represented predictiveness and displayed increased responses when outcome predictions became more reliable. These results extend previous findings on PH-like mechanisms in the amygdala and provide unique insights into human amygdala circuits during associative learning.

A Stress-Induced Shift From Trace to Delay Conditioning Depends on the Mineralocorticoid Receptor

“…In line with earlier studies, we focused on transient, learning-related activity (31)(32)(33), which is supposed to decrease as soon as the associations are learned (31)(32)(33)(34) and the US is reliably predicted (35)(36)(37). We expected learning-related activity in the amygdala for delay conditioning and in the hippocampus for trace conditioning.…”

Striatal structure and function predict individual biases in learning to avoid pain

“…The algorithm also accounts for each participant's baseline propensity to take gambles. Learning in the favored algorithm is weighted by associability (10,11), and the choices made for each deck tend to persist (7) (see SI Appendix for details of all learning algorithms and a validation of the model comparison procedure).…”

Amygdala subnuclei response and connectivity during emotional processing

“…Previous high-resolution fMRI studies (Table 1) reported involvement of corticomedial and CeM AG in various aspects of Pavlovian conditioning and instrumental processes (Boll et al, 2013;Prévost et al, 2011Prévost et al, , 2012Prévost et al, , 2013. However, because prior high-resolution fMRI literature did not report CeM activations in response to aversive visual stimuli, we must turn to animal experiments to gain greater insight into biological significance of our results.…”

“…Furthermore, there are substantial variations in methodology used to localize intra-AG activations. Generally, two main references have been used: Mai (2008) anatomical atlas (Bach et al, 2011;Boll et al, 2011Boll et al, , 2013Gamer et al, 2010;Prévost et al, 2011Prévost et al, , 2012Prévost et al, , 2013 or Amunts et al (2005) probabilistic atlas of the human AG subnuclei (Ball et al 2007(Ball et al , 2009Frühholz and Grandjean, 2013;Grant et al, 2015;Roy et al, 2009;Styalidis et al, 2014).…”