Hot Publication - Dombrovski et al.
Corticostriatothalamic Reward Prediction Error Signals and Executive Control in Late-Life Depression
Dombrovski AY, Szanto K, Clark L, Aizenstein HJ, Chase HW, Reynolds CF and Siegle GJ
Psychological Medicine, Published Online
The vascular depression hypothesis posits that prefrontal disconnection by vascular lesions causes depression and executive dysfunction. In a study recently published in Psychological Medicine, Dr. Alexandre Dombrovski and his colleagues examined reinforcement learning in late-life depression in the theoretical framework of the vascular depression hypothesis.
A cardinal feature of depression is that the behavioral effectiveness of reinforcement is weakened or distorted. Learning theory suggests that the problem may lie in representing unexpected rewards that normally drive behavior. The behavioral effectiveness of a reward depends not only on its magnitude, but also on how surprising it is. The degree to which a reward is surprising or unexpected—and behaviorally effective—is the reward prediction error. Prediction error can be calculated as the discrepancy between the expected and actually received reward. In every episode of learning (e.g. trial), the reward expectancy is updated by the prediction error. Reward expectancy, in turn, controls approach behavior. Thus, an altered representation of prediction errors may explain why depressed patients struggle to establish rewarding routines. Reward prediction errors are thought to be signaled from the dopaminergic midbrain to the striatum. They influence behavior through synaptic modification of corticostriatothalamic circuits. Cortico-striato-thalamic reward prediction error signals appear disrupted in younger and older patients with major depression.
Dr. Dombrovski and colleagues have previously shown that prediction errors (PE) modulate a cortico-striato-thalamic network in healthy older adults and that this modulation is blunted in patients with late-life depression. What is the mechanism of this blunting? To test this prediction, Dr. Dombrovski’s fMRI study compared 31 patients aged 60 years and older with major depression to 16 controls. Using a computational model, the researchers estimated neural and behavioral responses to reinforcement in an uncertain, changing environment (probabilistic reversal learning).
One mechanistic account is offered by the vascular depression hypothesis. It proposes that a disconnection of prefrontal and cingulate cortex by vascular lesions, indexed by white-matter hyperintensities on MRI, leads to the dyad of depression and executive dysfunction. If a single lesion--prefrontal/cingulate disconnection--caused depression and executive dysfunction, we would expect both to explain shared variance in neural PE signals. To test this prediction, Dr. Dombrovski’s fMRI study compared 31 patients aged 60 years and older with major depression to 16 controls. Using a computational model, the researchers estimated neural and behavioral responses to reinforcement in an uncertain, changing environment (probabilistic reversal learning).
The researchers observed that both depression and executive dysfunction were strongly negatively related to the integrity of corticostriatothalamic prediction error signals. However, their effects were distinct, and they did not explain any shared variance in PE signals, making it unlikely that both were caused by the same lesion. White-matter hyperintensities independently predicted disrupted PE signals, but did not explain the effects of depression or executive dysfunction. Next, VD would predict that both depression and executive dysfunction would be related to disrupted functional connectivity between the striatum and its cortical targets. Such a boost in connectivity to positive PE was diminished in depressed subjects, but the effect of executive dysfunction was trivial. Finally, behavioral signatures of depression and executive dysfunction during reinforcement learning were dissociated in two independent samples. In summary, the data did not support the assertion of the vascular depression hypothesis that prefrontal disconnection by vascular lesions causes depression and executive dysfunction. Rather, functional connectivity and behavioral analyses point to a disruption of ascending mesostriatocortical reward signals in late-life depression – a hypothesis that needs to be tested in future studies.
Alexandre Y. Dombrovski, MD, Katalin Szanto, MD, Howard J. Aizenstein, MD, PhD, Henry Chase, PhD, Charles F. Reynolds III, MD, Greg J. Siegle, PhD (Department of Psychiatry, University of Pittsburgh)
Luke Clark, PhD (Department of Psychology, University of British Columbia, Vancouver, Canada)