Sensory preconditioning is a phenomenon of classical conditioning that demonstrates learning of an association between two stimuli during an initial phase where the two stimuli (S1 and S2) are presented together but never followed by reinforcement. During the first stage of the experiment, presentations of S1 are followed promptly by S2. In the second stage of the experiment, S2 is repeatedly paired with an unconditioned stimulus so that it becomes a conditioned stimulus. Finally, in the third stage of the experiment, S1 is presented by itself. The phenomenon is demonstrated if S1 elicits a conditioned response during stage 3. Parallels can be drawn between Sensory preconditioning and second-order conditioning and together these effects can be described by the term "Higher-order conditioning".
The term "Sensory preconditioning" was coined by W.J.Brogden in 1939 at John Hopkins University. The discover of this effect produces a problem for the Stimulus-Response (S-R) account of learning. It appears that in a sensory preconditioning procedure, the organism has learnt about the co-occurrence of two stimuli in the absence of a response. During the first stage of a sensory preconditioning procedure two neutral stimuli (CS1 and CS2) are paired together in an association either simultaneously or serially. However, these stimuli are neutral and do not alone produce a response. During stage two, the traditional CS1-->UCS association is established. Consequently, the CS2 that has never been directly paired with the UCS begins to elicit a conditioned response in subsequent test phases. This suggests that during the first stage S-S learning has occurred because the response cannot be elicited until the second stage.
Three mechanism are thought to gives rise to the effect,
1) S-S pathway of associations
2) Mediated Conditioning (S-UCs)
3) Gestalt Conditioning (associative)
A forward conditioning experiment using a between-subjects design, followed by CS1 extinction suggests the possibility of an S-S pathway (Rizley and Rescorla, 1972). The experiment used a conditioned suppression paradigm following an A--> X | X-->UCS| A-->CR design.
Stage one: Tone (CS2) was presented serially with a Light (CS1)
Stage two: Light (CS1) was conditioned to an UCS (shock) to establish first-order conditioning
Test: CS2 elicited a conditioned response seen in stage two as elicited by the CS1
Following this sensory preconditioning procedure, responding to the CS1 (Light) was extinguished by presenting the light in the absence of the shock. In a second-order procedure, this CS1 extinction did not affect the responding to CS2. However, in this sensory preconditioning experiment the extinction of CS1 transferred to CS2, suggesting an associative chain explanation whereby CS2-->CS1-->UCS-->Response.
For the tone to be paired with the response, it would need exposure to the shock during stage 1, but it doesn't. The response cannot occur until the second stage when the shock is presented. Therefore, an S-R account can be discarded as there is no response in stage one.