Scientists say they've uncovered striking new evidence of how alcohol addiction impacts the brain's learning systems - and how those systems may slowly adapt during recovery - in a new study published on 22/05/26.
Led by The University of Manchester and The University of Huddersfield, they combined traditional EEG brain‑wave analysis with cutting‑edge machine‑learning tools to probe how people with a history of alcohol dependence learn from rewards and punishments.
The researchers used a reward-learning game - which they asked 20 abstinent alcohol-dependent and 26 healthy volunteers to complete while their brain activity was recorded.
The team found that both groups performed the task just as well as each other, however their brain signals told a different story.
A key brain response called feedback‑related negativity (FRN)- which reflects how we react to mistakes or bad outcomes - was reduced in people with a history of alcohol dependence.
This blunted signal appeared after both good and bad outcomes and did not vary with how long someone had been abstaining from alcohol.
The scientists say this could be a stable trait of alcohol dependence, reflecting underlying reward processing differences in people who are at risk of alcohol problems.
The study also looked at another signal, the feedback‑P3, which shows how strongly your mind reacts when you get important feedback and starts updating what you've learned.
Overall, it did not differ between the groups, but for people recovering from alcohol dependence, this signal was largest in the early stages of abstinence, and after many years appeared more similar to that of healthy people.
Researchers say this may reflect a brain change linked to abstinence itself.
To dig deeper, the team used a machine learning method called tensor decomposition to uncover hidden patterns in the EEG signals.
In the people with alcohol dependence, this revealed unusually early and strong activity in centro‑frontal brain regions near the top and front of the head.
This early surge was most pronounced in those in the earlier stages of recovery and could reflect, the scientists say, heightened sensitivity to feedback or a compensatory mechanism helping people maintain performance despite alcohol‑related brain changes.
They also found that healthy volunteers showed a different pattern, with a later burst of activity in the brain's parietal lobe, linked to processing sensory information before evaluating reward value.
The researchers used unsupervised machine learning - a method that finds patterns without being told what to look for - to break down the large amounts of EEG data.
This helped discover overlapping brain signals would have been difficult to spot using traditional methods alone.
Lead author Dr Mica Komarnyckyj from The University of Manchester, who is funded by the National Institute for Health and Care Research (NIHR) Manchester Biomedical Research Centre (BRC) as part of its Mental Health Theme , said: "Alcohol dependency is a complex and challenging health condition, and many people have difficulties maintaining recovery despite treatment and support.
"We believe our findings offer fresh insight into how alcohol dependence can influence the brain systems involved in learning and reward.
"Larger, long‑term studies are now needed to understand if the EEG markers we identified here could one day help track recovery or identify those people who might need extra support.
Researchers conducting the study are funded by the is UKRI Future Leaders Fund, the Biotechnology and Biological Sciences Research Council, and the National Institute for health and Care Research (NIHR) Manchester Biomedical Research Centre. It is published in the journal Clinical Neurophysiology.
- The paper Altered EEG markers of reward learning during abstinence in alcohol dependence: a probabilistic reversal learning study is available hereDOI
