Blocking, Overshadowing and Related Concepts. Fig. 3.
Overshadowing shown as reduced learning relative to the conditioning control group in a standard fear conditioning (lick suppression) procedure. The dependent variable is the log transform of the time taken to complete 10 licks in the presence of the CS. Overshadowing was shown in the saline-(SAL) and SCH23390- (SCH, dopamine D1-like antagonist) injected groups. Overshadowing was abolished by treatment with amphetamine (AMP) and the amphetamine-induced abolition of overshadowing was reversed by treatment with SCH23390, suggesting that dopamine D1-like receptors mediate the amphetamine-induced abolition of overshadowing. (Adapted from O'Tuathaigh and Moran (2002) Psychopharmacology 162:225-231, see text for full details.)
blocking, overshadowing, and related effects. To date, the role of dopamine has been an understandable focus for pharmacological studies, principally of latent inhibition, though also of blocking and other selective learning effects, due to the fact that established ► antipsychotics are dopamine antagonists. The modulation of these effects by other neurotransmitter systems remains to be established. For example, based on the known effects of serotonergic treatments on latent inhibition (Weiner 2003), ► serotonin would similarly be expected to modulate blocking and related effects.
The blocking effect has been targeted, to understand the biological basis of attentional abnormality, to a large extent without a full consideration of its underlying psychological mechanisms. This situation may soon be remedied in that blocking has recently assumed prominence as a tool to investigate the neural basis of
Blocking, Overshadowing and Related Concepts. Fig. 4.
Blocking shown as a low percentage of baseline learning in a contextual fear conditioning procedure. The comparison group controls for overshadowing by context which it can be seen was unaffected by any of the drug treatments. Blocking was increased by injection of amphetamine in n.acc and decreased by combined D1 and D2 receptor antagonism produced either by injection of cis-(z)-flupenthixol or combined treatment with SCH23390 and sulpiride. In isolation, SCH23390 and sulpiride were without effect, suggesting that a combined action at dopamine D1-like and D2-like receptors mediates the amphetamine-induced abolition of blocking. (Adapted from lordanova et al. (2006a) Eur J Neurosci 24:3265-3270, see text for full details.)
► Prediction Error. Prediction error is fundamental to normal associative learning because it provides the basis on which to learn when there is a discrepancy between what is expected, and what actually occurs. Thus, uncertainty is an important factor in determining attention in learning and dominant learning theories propose that new learning requires that an association is not already at full strength. Positive prediction error generates excitatory conditioning; negative prediction error generates inhibitory conditioning (Fig. 5).
Electrophysiological studies show the pivotal role of
► dopamine as a neurochemical substrate of prediction error (Schultz 2006; Schultz and Dickinson 2000). Blocking was the key paradigm that drove our understanding of associative learning in terms of prediction error: the Stage 1 pre-training means that the UCS is fully predicted by the time that the competing cue is introduced in Stage 2. Thus, there can be no prediction error and thus no additional conditioning to the redundant cue. However, prediction error can be systematically manipulated by changing the UCS delivered. If the UCS is other than expected, ► Unblocking occurs. This can take the form of additional excitatory conditioning to the additional CS
if the UCS is more than expected (e.g., a higher intensity foot shock is delivered in upshift unblocking) or inhibitory conditioning to the additional CS if the UCS is less than expected (e.g., a lower intensity of foot shock is delivered in downshift unblocking). Electrophysiological studies have confirmed that some of the same populations of dopamine neurons that show increased activation following the presentation of excitatory CSs, that signal ''more than expected,'' show depressed neuronal firing following the presentation of inhibitory CSs, that signal ''less than expected'' (Schultz 2006). These neurons show no change in neuronal firing when there is zero prediction
Blocking, Overshadowing and Related Concepts. Fig. 5.
Positive prediction error when the probability (p) of the UCS is increased on presentation of the CS in question generates excitatory conditioning. Negative prediction error when the probability of the UCS is decreased on presentation of the CS in question generates inhibitory conditioning. When p(UCS/CS) = p(UCS/no CS) along the diagonal trend line, there can be no new learning.
error, in other words when things remain ''as expected.'' Such a case is provided in the blocking procedure. Pre-training with CS A normally blocks learning about CS B because the prediction error is small. The standard blocking procedure is readily adapted to study the neural bases of excitatory and inhibitory learning (Fig. 5) by manipulating the UCS delivered (Table 2).
Thus, blocking and unblocking variants provide target tasks to identify the neuropharmacological substrates of prediction error, for example, using microinjections into the nucleus accumbens (Fig. 6).
Thus, analysis of blocking from an associative learning theory perspective has identified the same underlying neural substrates that are the target of translational studies of blocking as an endophenotype for schizophrenia. At the behavioral level, abnormalities in the processing of prediction error may be the cause of the formation of inappropriate associations in schizophrenia. In other words, studies of prediction error, known from electro-physiological studies to depend on the dopamine system, also further our understanding of dopaminergic disorders such as schizophrenia. Deficits in blocking, seen both in schizophrenia and under amphetamine, represent a paradigm instantiation of abnormalities in the processing of prediction error. Future studies of the relevant neural substrates of prediction error should include, but are not restricted to, those identified in blocking procedures. Downstream from these Pavlovian effects, abnormal processing of prediction error has been linked to abnormalities of action, including drug addiction where overconditioning in hyperdopaminergic states could promote cue-driven relapse because of the increased representation of drug-related cues (Montague et al. 2004; Schultz and Dickinson 2000).
Advantages and Limitations of Blocking and Overshadowing
Latent inhibition, now well established as a model for schizophrenia, shows the predicted sensitivity to psycho-active drugs, and human participants with schizophrenia
Was this article helpful?