The Question: What Is a Reward Worth Once You Have to Wait?
Delay discounting asks how the value of a reward falls as the cost of waiting rises — the core of temporal decision-making and a classic impulsivity axis. At NEATLabs, rats made these trade-offs while we asked whether a distributed electrophysiological signal tracks subjective reward value.
behavioral paradigm map
Delay Discounting — reward value & subjective value
Rats choose between smaller-sooner and larger-later rewards while reward-locked beta oscillations track reward magnitude, delay cost, and modeled subjective value.
Task design
Choice options
Smaller-sooner vs larger-later
Delay manipulation
Increasing waiting cost
Reward delivery
Magnitude and timing vary
Neural measurement
Reward-locked LFP
Time-frequency around reward
Reward network
OFC, mPFC, insula, vStr, amygdala
Subjective-value model
Computational value estimates
What we found
Beta = magnitude
Reward-locked beta scales with reward size
Delay decay
Beta power decays with longer delays
Value correlate
Beta tracks modeled subjective value
Study Design and Rat Behavior
The task pits smaller-sooner against larger-later rewards and manipulates the delay to the larger option. The behavior it elucidates:
- Preference curves — how strongly an animal discounts future reward as delay grows.
- Sensitivity to magnitude — whether a bigger reward pulls choice even under a longer wait.
- Individual impulsivity — steeper discounting means a more impulsive animal.
Each choice, delay, and reward-delivery moment had to become a clean event so that neural activity could be locked precisely to reward receipt.
Methodology
- Recording: rodent local field potential recordings across the reward network — orbitofrontal cortex, medial prefrontal cortex, anterior insula, ventral striatum, and amygdala.
- Analysis: reward-locked time-frequency analysis, isolating oscillatory power evoked at the moment of reward across delays and magnitudes.
- Modeling: a computational model of subjective value, so neural signals could be tested against a per-trial estimate of what the reward was worth to the animal, not just its objective size.
What We Found
The paper identified cortico-striatal beta oscillations as a reward-related signal:
- Reward-locked beta power signaled reward magnitude and decayed with longer delays — moving with value, not just sensory reward.
- OFC, mPFC, anterior insula, ventral striatum, and amygdala electrodes all showed beta-frequency reward signals, making this a genuinely distributed network signal.
- Beta power correlated with subjective-value estimates from the computational model — tying the physiological signal to a behavioral, model-based quantity.
Publications
- Cortico-striatal beta oscillations as a reward-related signal — Cognitive, Affective, & Behavioral Neuroscience (2024)
- Cortico-Striatal Beta-Oscillations as a Marker of Learned Reward Value — bioRxiv preprint (2022)
This paradigm is one thread of the broader NEATLabs research program.