Unanticipated contradiction between percepts or thoughts and the predictions of organized knowledge

The trouble with political jokes is that they often get elected. For most if not all the readers, encountering this statement here is a real surprise. This is precisely why it stands a good chance to be remembered. Surprise ('sur'- + 'prendre', Old French, prehendere, 'to seize' in Latin) is a perceptual "dimension well known to affect the "acquisition of memory in "real-life; it may lead, for example, to an enduring "'flashbulb memory'. Descartes regarded 'wonder', or 'sudden surprise', as one of only six primitive 'passions of the soul', and appreciated its role as a trigger of "attention and learning: 'It has two causes: first, an impression in the brain, which represents the object as something unusual and consequently worthy of special consideration; and secondly, a movement of the spirits, which the impression disposes both to flow with greater force to the place in the brain where it is located so as to strengthen and preserve it there...' (Descartes 1649). Note, by the way, how well this paragraph fits five centuries later into discussions of "functional neuroimaging; all you have to do is to replace 'spirits' with 'blood', but this deserves a separate discussion on the history of ideas. A more recent forefather of our scientific "zeitgeist, Darwin (1872), also appreciated the importance of surprise as a primitive, universal emotion, characterized by specific gestures and physiological response. He focused on unpleasant surprises, startles, and fears, whereas we all know that benevolent surprises do exist, albeit they do not show up too often. In contemporary neuroscience, 'surprise' and its roles in learning and memory can be discussed at multiple "levels, from the behavioural to the molecular and vice versa. Let's therefore look at some of its manifestations in the brain, neuronal circuits and individual neurons.

'Surprise' is basically a sudden, significant mismatch between the actual and the expected. In brains, this is between on-line inputs (be them sensory "percepts or endogenous "internal representations) and off-line internal representations, i.e. memories. Definition 2 above emphasizes two additional properties: (a) that the knowledge is organized, and not merely a collection of data, and (b) that such organized knowledge makes predictions about reality ("a priori, "planning). These properties are explicit in the terminology of cognitive psychology. In cognitive terminology, 'surprise' is a sudden discrepancy between input and a 'cognitive schema'. 'Schemata' are structured clusters of generic knowledge, that represent situations, events, actions, or complex objects, enable the comprehension of input, and predict future outcome of action (Eyesenck and Keane 1995; those schemata that contain organized sequences of stereotypical actions are 'scripts'). When external or internal data suddenly contradict the prediction of a schema, surprise follows. This may then motivate and enable the analysis of the discrepancy and the adjustment to it (Schutzwohl 1998). Thus, in a way, surprise is a sudden perturbation of cognitive "homeostasis.

Although the aforementioned framework connotes human or at least primate cognition, there is no reason why the basic elements should not be adapted to "reductive treatment of much simpler brains. Conditioning paradigms were particularly instrumental in casting light on the postulated role of surprise in learning in a variety of species. Consider, for example, the phenomenon called 'blocking', which is the inhibition of the conditioning to a stimulus, CSp in a compound CS1 + CS2 stimulus, by previous pairing of CS2 with the UCS ("classical conditioning). A prevalent interpretation of the effect is that in order for an association between a CS and a UCS to be formed, the UCS must surprise the animal, but in blocking this is not the case, as CS2 already predicts the UCS (Kamin 1969; for alternative interpretations see Mackintosh 1983). The analysis of conditioning has provided incentives as well as constraints for a number of formal "models of learning. Among them is the noted Rescorla-Wagner "algorithm (Rescorla and Wagner 1972), which, in a nutshell, concludes that the amount of learning is proportional to the amount of surprise. In other words, learning theories, not only layperson intuition, also mark surprise as a driving force in learning.

Recording the electrical activity of the human brain in action by electroencephalography (EEG, "functional neuroimaging), has identified brain waves that appear only under 'surprising' situations. The EEG of individuals that respond to a rare stimulus occurring randomly in a sequence of frequent stimuli, or to an omission of an expected stimulus, shows a characteristic wave about 300 ms after the surprising event ('P300 wave'; Sutton et al. 1965). Similarly, a characteristic evoked-response brain wave about 400 ms after the stimulus is detected in individuals that encounter an out-of-context word in a sentence reading task ('N400 wave'; Kutas and Hillyard 1980). These are striking physiological correlates that differentiate fast cognitive responses by time and type.

What are the brain circuits involved? In theory, one expects circuits that compare on-line with off-line representations, identify the mismatches, induce other circuits to generate the proper behavioural response, and trigger the proper long-lasting representational change. To be useful ecologically, these circuits must operate in the subsecond range, even if they process input modalities that are considered relatively slow, such as taste (Halpern and Tapper 1971). Plausible candidates are systems that involve "cortex (both frontal and modality specific), thalamocortical, or thalamo-cortical-brainstem circuits. Consider, for example, the following candidate scheme: on-line information is encoded in the brainstem or cortex or both, off-line information in the cortex, and the thalamus does the comparisons (Ahissar et al. 1997). Furthermore, there are good reasons to assume that the match/mismatch output signal modulates diffused neuromodulatory systems, such as the cholinergic ("acetylcholine; e.g. Mishkin and Murray 1994; Naor and Dudai 1996), "dopaminergic (e.g. Schultz et al. 1997; Redgrave et al. 1999), or "noradrenergic (e.g. Kitchinga et al. 1997). These neuromodulators are then expected to regulate "intracellular signal transduction cascades in the target neurons, culminating in "synaptic remodelling and ultimately in long-term memory (Berman et al. 1998). Thus at the cellular level, the mechanisms that encode surprise merge with those that encode attention and subserve acquisition of memory. If we delve into the nuts and bolts of these signal transduction cascades, we could even end up with molecular models that account for the ability of surprising information to encode a robust "engram after only a single brief experience, by shifting instantly the balance of the signal transduc-tion cascades and the transcription factors in favour of that configuration that activates the appropriate 'long-term "plasticity genes' (Bartsch et al. 1995; "CREB, "immediate early genes, "fear conditioning, "flashbulb memory).

It is noteworthy that some of the aforementioned studies, especially those involving cellular and molecular analysis, do not target 'surprise' specifically, but rather the reaction to unfamiliar events in general. This should not be taken to imply that unfamiliarity and surprise are utterly identical. Unfamiliarity can be detected by any sensory system with an access to memory; surprise as defined here requires in addition the ability to generate expectations on the basis of organized knowledge. In addition, novelty is a continuous dimension (many inputs are only slightly novel), whereas bona fide surprise is probably more of an abrupt event, conforming to a step function. Still, it is likely that brain mechanisms that subserve detection of surprise overlap with those that subserve the response to unfamiliarity in general. In all theses cases, the brain compares the present to the past. In the case of surprise, an additional faculty is engaged, that of computing expectations and evaluating the significance of their sudden clash with reality, in real-time. We do not know whether even the simplest nervous systems that can detect novelty generate such rudimentary expectations and are engaged in comparison of pre-representations with on-line input ("a priori). In mammals, this ability is contributed, among others, by the frontal cortex (Fuster 1995a; Watanabe 1996; Daffner et al. 2000; "attention, "working memory).

Surprises are effective incentives for behavioural modifications that contribute to survival. This is probably why the ability to detect them had been embedded effectively in our brain. Seen this way, the most important ones are the bad surprises, because they may kill; hence, Darwin was not off-track when he emphasized startling, nasty surprises (Darwin 1872; also 'startle reflex'under "attention, "sensitization). With the emergence of human "culture, our capacity to note mismatches with predictions of organized knowledge has gained additional functions, some of which are geared to create pleasure, not pain. Hence, it is said that surprise and 'defamiliarization' are at the basis of our appreciation of art (Shklovsky 1917). On the more practical side of life, surprises, although still mostly bad ones, contribute to our attitudes as consumers of goods (Maute and Dube 1999). And in the small world of universities, professors can exploit pedagogical surprises to enhance the success of their classes (Kintch and Bates 1977; Thorne 1999).

Selected associations: A Priori, Algorithm, Attention, Dimension, Sensitization

Living Life

Living Life

Want to be and stay positive? Download This Guide And Discover 50 Tips To Have A Positive Outlook In Life. Finally! Reach your goals, face your fears and have a more fulfilled and happy life.

Get My Free Ebook


Post a comment