Changes in emotional processing precede and predict clinical benefit

Many patients have both psychological and pharmacological treatments. Rather than being opposing philosophies of intervention, psychological concepts show how effects on neurocircuitry translate into clinical benefit. The way antidepressants reduce negative bias in the processing of emotional cues seems central. Catherine Harmer, Professor of Cognitive Neurosciences at the University of Oxford, UK, has been closely involved in this work.¹ 

Empirically, people prone to depression seem more likely than others to focus on a negative interpretation of events and to show a negative bias in attention and memory. There is increasing evidence from controlled studies that such bias in emotional processing is not simply a correlate of low mood but an aetiological factor and – most importantly – a factor that responds to treatment.

Work with patients suffering from depression shows that antidepressants quickly reverse negative bias in the processing of emotional stimuli.² Such effects are seen far before any clinical benefit is apparent and can be interpreted as restoring patients’ capacity to experience the world in a more positive way.

Changes in response to emotional stimuli early in the course of antidepressant therapy relate to changes seen in functional MRI. In the individual patient, absence of an early reduction in negative bias may in the future become established as a robust enough predictor of lack of clinical efficacy to justify switching drugs³, according to Professor Harmer. This hypothesis is now under investigation. More generally, change in emotional processing may be a way of screening new compounds for potential antidepressant action.⁴

Psychological concepts help us understand how antidepressants work

Core psychological processes

Studies of emotional memory show that significantly reduced recall of positive terms relates to the severity of subsequent depression. Patients with depression also show a reduced capacity to recognise faces expressing positive emotion. Such research has consistently suggested that negative bias in the processing of emotionally significant stimuli is among the psychological mechanisms central to initiating and maintaining the condition.

Importantly in practical terms, we now know that this bias can be countered by antidepressant therapy.

In a study of healthy controls and patients with depression, administration of a single dose of antidepressant was sufficient to bring patients’ capacity to recognize happy faces close to that of controls.⁵ This effect was not seen with placebo. A study involving memory for words with positive emotional connotation showed that antidepressant treatment achieved a similar normalisation of performance.

So it seems that pharmacological therapy can very rapidly set the scene for patients to start perceiving the world more positively. This may pave the way to therapeutic response -- even though it may be weeks before patients start to feel better and before rating scales pick up measurable improvements in symptoms.

Processing changes predict clinical response

Evidence that effects on processing bias can predict therapeutic efficacy, which has now been independently replicated, came initially from collaborative work involving Catherine Harmer in Oxford and Richard Tranter at the University of Bangor, North Wales.⁶

Patients whose processing of happy facial expressions had improved two weeks after starting antidepressants were more likely than those showing little change to experience clinical improvement at six weeks.  

So such early improvement may be a necessary step towards therapeutic response. But it may not be sufficient.

Even if the bias becomes less negative, there have to be positive experiences to be processed. Work on self-rated levels of interpersonal support suggests that patients in unremittingly adverse environments do not experience improved mood despite re-setting of their processing capacity by treatment.

Hence understanding the therapeutic response to antidepressants requires that we acknowledge the interplay of three factors: psychopharmacology, psychological processing, and social and environmental context.

Antidepressants affect emotional processing before a patient notices an improvement in how they feel

Correlates in brain function

The corticolimbic network, and particularly the amygdala, is thought to be important in determining emotional responsiveness. Functional MRI studies have found a relationship between vulnerability to depression and high responsiveness to negative stimuli (relative to positive ones) in the amygdala. This suggests that the brain in depression is treating negative experiences as more deserving of attentional resources than positive ones.

But it seems that this bias towards negative facial expressions seen on fMRI can be reduced by seven days’ treatment with an antidepressant -- even though depressive symptoms at that stage have not been lessened.

Early effects on emotional processing distinguish between patients who will respond to antidepressants and those who will not

Processing bias as a biomarker in drug screening

In developing new antidepressants, much reliance is placed on animal models. But these have low predictive validity. So there is growing interest in using studies of drug effects on emotional processing as a marker of potential clinical efficacy.³

In a recent study of a novel agent in development, an eight week randomized clinical trial showed only marginal benefit relative to placebo. At one week, this agent had shown similarly marginal effects on emotional processing. This absence of early effect might have been used as a reason not to pursue the study.

That case is strengthened by the fact that the subgroup of patients who did ultimately benefit from active treatment had shown improved recognition of happy faces at week one. The placebo group did not show this association, suggesting that change in recognition bias for facial emotions was specific in picking up a drug effect.

Change in emotional bias to guide treatment

Characterisation of the mechanisms underlying the dynamics of emotional processing in depression may eventually lead to better decisions on treatment. However, the empirical value of the correlation between emotional bias and overall clinical effects remains to be proven and requires an integrated clinical research approach. This is now being undertaken.

In the European Union-funded PREDICT study, which starts later in 2016, primary care patients with depression needing an SSRI will be randomised to treatment as usual or treatment guided by studies of emotional processing.