Ketamine Restored Reward Bias in Depression and Stressed Rats

TL;DR: Ketamine lifts depression scores fast, but anhedonia — the inability to learn from positive outcomes — is what disables many patients. This translational study put the same probabilistic reward task in front of treatment-resistant depression patients and chronically stressed rats. Within 24 hours, ketamine restored reward bias to healthy-control levels in both species. General task discriminability did not change.

Source: Biological Psychiatry Global Open Science (2026) | Bogdanov et al.

Ketamine can lift depressive symptoms quickly, but many patients are most disabled by anhedonia — the loss of reward, motivation, and the ability to learn from positive outcomes. This study tested that symptom domain directly with a probabilistic reward task designed to work in both humans and rats — and found ketamine moving reward bias the same way in both species.

Why Anhedonia Needed a Cross-Species Behavioral Test

Ketamine research has had a translation problem. Rodent studies identify mechanisms, but the behavioral readout often fails to match what patients actually describe. The probabilistic reward task helps close that gap. It measures whether a participant develops a response bias toward the stimulus rewarded more often — testing whether the nervous system is starting to favor the option that has been paying off.

The task is relevant to anhedonia because reward learning can be measured without asking someone to summarize their mood. It also gives animal work a behavioral bridge to human depression that is more functional than rote symptom-checklist analogs.

The Human and Rat Tasks Pointed the Same Way

The study tested people with treatment-resistant depression and rats exposed to chronic stress. Human ketamine dose: 0.5 mg/kg. Rat dose: 10 mg/kg. The point was not to treat those doses as directly equivalent but to ask whether the same behavioral signature moved in the same direction.

Ketamine increased response bias toward the more frequently rewarded stimulus in both species. By 24 hours after administration, reward bias reached levels comparable with healthy controls. The 24-hour window is part of the signal — ketamine’s clinical effects appear before slower antidepressant mechanisms would be expected, and a reward-learning shift inside that window supports the idea that rapid synaptic or circuit-level changes can alter how positive feedback shapes behavior.

The Discriminability Null Was the Important One

The negative finding does the heavy lifting. Ketamine did not appear to make participants or rats simply better at the whole task. It selectively changed the reward-learning component. That distinction protects the interpretation. A drug that improves attention, speed, or compliance would inflate task performance without necessarily touching anhedonia. Here, the result points more directly at reward responsiveness:

• Reward bias changed: response bias shifted toward the more frequently rewarded stimulus.
• Discriminability did not: the participants and rats were not simply better at telling stimuli apart.
• Anhedonia subgroup mattered: exploratory analyses suggested the effect was strongest in people with more pronounced baseline anhedonia.

Why This Reframes Ketamine’s Antidepressant Effect

The paper does not identify the full molecular mechanism. It supports a behavioral mechanism: ketamine can restore the brain’s ability to update choices from positive feedback. That is more specific than saying ketamine reduces depression scores. It points to a patient-relevant question — does the drug reopen a window where reward becomes learnable again?

Reward bias narrows the depression question helpfully. Depression rating scales mix many symptoms into one score. Sleep, appetite, guilt, anxiety, psychomotor change, sadness, and suicidality can all move together or apart. Anhedonia can remain even when total depression scores improve. This task isolates one piece: the ability to update behavior from reward. That is not the whole of depression, but it maps onto a problem patients recognize immediately — good outcomes stop pulling behavior forward.

What the Cross-Species Match Buys Drug Development

Many antidepressant mechanisms look elegant in animals and blurry in humans. The advantage here is that the investigators used functionally analogous tasks. The human and rat data are not identical, but they point at the same behavioral construct. If a future compound changes the same probabilistic reward readout in rodents, researchers have a stronger reason to test it for anhedonia in patients. The task becomes a bridge rather than a loose metaphor.

The study also points to a possible clinical subgroup. If the ketamine effect is strongest in people with higher baseline anhedonia, reward-task measures become a logical way to identify who is most likely to benefit. Treatment-resistant depression is not one biology — a patient dominated by anhedonia needs a different readout from a patient whose main burden is anxiety, rumination, insomnia, or suicidal crisis.

What This Study Cannot Tell Patients Yet

The paper does not show how long the reward-bias effect lasts, whether it tracks subjective pleasure, or whether repeated ketamine keeps improving reward learning. It does not prove that reward bias is the main driver of clinical response. A stronger clinical study would pair reward-task changes with anhedonia scales, daily-life motivation measures, and longer follow-up — testing whether the laboratory reward signal predicts the part of depression patients actually want back: the ability to feel drawn toward life again.

If reward bias tracks ketamine’s anhedonia effect, it becomes more than a laboratory outcome. It gives researchers a way to compare dosing schedules, adjunctive psychotherapy, or drugs meant to prolong ketamine’s benefits. The strongest version of this finding would show that early reward-bias improvement predicts later gains in motivation, pleasure, and real-world activity. That is the bridge from task performance to clinical meaning — and this paper builds the first useful span of it.