Reward Prediction Error, Dopamine, and Why Unpredictable Relationships Feel Addictive

In neuroscience, one of the most important mechanisms behind motivation, desire, and attachment is Reward Prediction Error (RPE), a dopamine-based learning signal discovered in the work of neuroscientist Wolfram Schultz and later applied to human behavior, addiction, and relationships.

What is Reward Prediction Error?

Reward Prediction Error describes the gap between what we expect and what we actually receive.Dopamine doesn’t simply rise when something feels good — it rises most strongly when the reward is better than expected.

• Positive RPE: Reward is greater than expected → dopamine spikes

• Zero RPE: Reward matches expectation → stable dopamine

• Negative RPE: Reward is worse than expected → dopamine drops

This mechanism was observed in classic neuroscience experiments where dopamine neurons fired more intensely not after the reward, but when the reward was unpredictable.(Source: Schultz, W. (1997). Neuron, “A Neural Substrate of Prediction and Reward.”)

Dopamine = Desire, Not Pleasure

A key point: Dopamine is not about pleasure. It is about anticipation, craving, and motivation. This is why unpredictability is so powerful. When rewards (love, affection, attention) come sporadically or inconsistently, the brain generates stronger dopamine spikes, because each positive moment becomes a “surprise reward.”

This creates a loop very similar to gambling:

• You never know when the “reward” is coming →

• Your brain stays highly alert and invested →

• The unpredictability increases desire

• Desire keeps you attached, even when the relationship is unhealthy

Why Unhealthy Relationships Become Addictive

Unpredictable partners emotionally unavailable, inconsistent, hot-and-cold create the perfect neurochemical environment for addictive bonding. This is sometimes called intermittent reinforcement, a concept from behavioral psychology (Ferster & Skinner, 1957) and heavily researched in addiction studies. Inconsistent rewards are the strongest driver of attachment.

Examples in relationships:

• One day: warm, loving, apologetic

• Next day: distant, critical, withdrawn

• Then: sudden beautiful moment, affection, or connection

Each unexpected positive moment creates a positive Reward Prediction Error, triggering a bigger dopamine spike than consistent affection ever would. The nervous system learns: “Maybe this time it will be good again.” This keeps people trying harder, waiting longer, tolerating more pain, and hoping for the next “reward.”

Why many people ignore red flags

From a psychological perspective:

1. The nervous system becomes conditioned to expect rare highs

2. Dopamine reinforces the chase, not the connection

3. Early attachment patterns (e.g., unpredictable caregiving in childhood) make unpredictability feel familiar

4. Surprise affection activates reward circuits much more intensely than stable love

5. The person becomes emotionally dependent, not emotionally connected

This is why predictable partners often feel “boring,” while chaotic ones feel “intense.” The intensity is not love it is dopamine responding to unpredictability.

Clinical and trauma-informed perspective

People who grew up with:

• emotionally unavailable parents

• inconsistent caregivers

• unpredictable love or validation

• trauma bonds often have a nervous system wired to seek familiar patterns, not healthy ones. Your brain isn’t addicted to the person, it is addicted to the intermittent reward cycle.

As a psychotraumatologist, I can recommend:

• Recognizing how trauma-shaped nervous systems confuse intensity with intimacy

• Understanding that desire does not always reflect what is good for you it reflects what your brain expects

• Working on emotional regulation and attachment wounds to rewire what feels “normal” and “safe”

• Choosing relationships where emotional presence is consistent, not unpredictable

Healing requires teaching the nervous system that stable love is exciting, not dull.

Suggested sources for credibility

• Schultz, W. (1997). A Neural Substrate of Prediction and Reward. Neuron.

• Schultz, W., Dayan, P., & Montague, P. (1997). A Neural Substrate of Prediction and Reward. Science.

• Berridge, K. (2007). The debate over dopamine’s role in reward: “liking” vs “wanting.” Psychopharmacology.

• Ferster, C., & Skinner, B. (1957). Schedules of Reinforcement.

• Panksepp, J. (1998). Affective Neuroscience.