The idea being tested is how reinforcement schedules depend on the number of responses rather than time. A variable ratio schedule delivers reinforcement after an unpredictable number of responses, with the average around a certain value but the exact count varying each time. This unpredictability makes behavior very persistent and high in frequency, because the next reward could come at any moment, so the learner keeps responding to maximize chances of reinforcement. A gambling-like pattern like slot machines is a classic example: you don’t know when the next win will come, so you keep pulling the lever. In contrast, a fixed-ratio schedule gives reinforcement after a predictable number of responses, a fixed-interval schedule reinforces after a set amount of time (regardless of responses), and continuous reinforcement delivers a reward after every response. The unpredictable-number-of-responses reinforcement fits the variable ratio schedule best.

Reinforcement schedules tell you when a behavior will be rewarded. A set-number-of-responses schedule is fixed-ratio. The reinforcement comes after a specific, unchanging count of responses, like every fifth response. This makes the subject work quickly and consistently because the payoff is predictable, so they keep responding to reach the next reinforcement. After each reward, there’s often a brief pause before starting up again, and the length of that pause grows with the required number of responses. This stands in contrast to other schedules: variable-ratio delivers reinforcement after an unpredictable number of responses, yielding very high and steady response rates with few pauses; continuous reinforcement rewards every single response, which quickly teaches the behavior but doesn’t last long if rewards stop; fixed-interval delivers reinforcement after a fixed amount of time, regardless of how many responses occur, leading to a pattern of increasing responses as the time for the next reinforcement approaches and a lull right after reinforcement.

In Pavlov's experiment, a neutral stimulus is something that initially does not elicit the response of interest. The meat powder naturally causes the dog to salivate, so it is the unconditioned stimulus and the salivation to the meat powder is the unconditioned response. The bell, on the other hand, starts off not causing salivation, making it a neutral stimulus. After repeatedly pairing the bell with the meat powder, the dog begins to salivate to the bell alone, so the bell becomes a conditioned stimulus and salivation to the bell becomes the conditioned response. The dog is just the subject involved in the experiment.

The brief electrical charge that travels along a neuron when it fires is called an action potential. This impulse is an all-or-none event: if the neuron reaches the threshold, voltage-gated channels open and a rapid depolarization occurs, followed by repolarization, producing a spike of about the same strength every time. Once it starts, the action potential propagates along the axon by triggering the next segment to depolarize, so the signal travels without diminishing its amplitude (myelination can speed this up through saltatory conduction). The frequency of these spikes, not their size, carries information about stimulus intensity. This is different from the resting potential, which is the baseline negative charge inside the neuron when it isn’t firing, and from membrane potential, which describes the electrical difference across the membrane at any moment. Neurotransmitter release happens at the synapse as a chemical consequence of the action potential reaching the terminal.