Outline of a theory of manic-depressive Illness

Positive feedback loops
    The most distinctive characteristic of manic-depressive illness (including bipolar disorder) is locking into the antithetical mental states of mania and depression, and alternation between them. The essence of mania appears to be overconfidence, and the essence of depression, lack of confidence (a fuller characterisation of these states will be considered further on).
    Locking into extreme states is also characteristic of an important building block of electronics, the bistable or flip-flop. What gives the bistable this characteristic is a positive feedback loop with greater than unity gain (what this means will be made clear shortly). (The bistable should not be confused with the astable, which involves a delayed negative feed-back loop.) An important point is that the positive feedback loop only produces the bistable effect if its gain is greater than unity. This may be made clearer by the following consideration of feedback in a public address (PA) system (clever readers should note the comment further on!).
    A PA system consists of a microphone connected via an amplifier to a loudspeaker, and a positive feedback loop is created because sound from the loudspeaker can itself be received by the microphone. In normal operation this poses no problem because the feedbacking sound from the loudspeaker is much less than the original sound of the announcer's voice; that is, there is less than unity gain. But if the amplifier volume control is turned up too far, and/or if the loudspeaker is placed too near to the microphone, then the well-known feedback effect soon occurs, namely a continuous howling or whistling noise. This is because any sound entering the microphone will now produce a louder feedbacked sound from the speaker, which subsequently produces an even louder sound, until the system reaches its maximum output capability. (It will be noted that the PA system seemingly produces only a monopolar effect rather than bipolar; this is because the opposite of a loud sound is an equally loud sound of reverse phase.) (Clever readers may have realised that the PA feedback phenomenon may be at least as accurately characterised as involving an astable, with delayed feedback of pressure change producing alternations of pressure; they are not incorrect, but the interpretation as a bistable with virtually instantaneous feedback producing locking into extreme loudness is just as valid and is in any case here given merely for purpose of illustration.)

Positive feedback in behaviour and the brain
    An argument will now be presented that normal humans have evolved a psychological positive feedback loop which gives the characteristics of bipolar disorder when its gain exceeds unity.
    For an animal, human or otherwise, to be excessively cautious would not be optimal from the point of view of natural selection: such an animal would miss valuable opportunities. On the other hand excessive confidence would lead to commission of actions better not performed, and to setbacks. So natural selection could select for an intermediate generally optimum level of cautiousness/confidence. But the resulting animal would be rather inflexible in this regard. Circumstances change, and sometimes there are more hazards than opportunities, and sometimes the converse. A useful indicator to the animal of whether to be relatively confident or cautious would be the degree of success or failure it encounters: when the going gets easy, it should ideally become more confident so as to avoid missing opportunities (and consequent losing out in the 'survival of the fittest'), but when the going gets hard it should become more cautious to avoid further misfortune.
    Thus natural selection would favour an innate mechanism that causes 'successes' (i.e., believed successes) to increase confidence, and 'failures' (i.e., believed failures) to decrease confidence, or in other words increase caution.
    We must now consider more closely the nature of confidence and caution (the terms are used here as antonyms of course). Consider the reaction of two individuals presented with the same risks and opportunities: The confident one would see more prospects of success, the cautious one more prospects of failure. Thus increased confidence involves increased perceptions of success, real or otherwise. And the converse with caution and failure.
    Thus results the positive feedback loop. Perceptions of successes increase confidence, and that confidence in turn tends to increase perceptions of successes, which in turn further increase confidence, and so on; and conversely with caution and failure.

Excessive feedback causing psychosis
    Up to a point these innate tendencies would be advantageous. But in the event that their magnitude is sufficient to produce greater than unity gain in the feedback loop, then the bistable effect would result, with strong tendency towards extreme overconfidence or extreme caution and locking into the extreme state. The individual would remain locked in the extreme state until occurrence of a change of internal or external factors such as to cause a reduction of the feedback. This reduction could be sustained, producing remission from psychosis; or it could be transient, immediately followed by a return to an extreme state.
    Whether a particular individual going above the unity gain threshold at a particular time would fall into overconfidence or into overcaution instead would depend on his circumstances (encouraging or discouraging) and perhaps on constitutional disposition. It seems likely that some individuals would by reason of circumstances and or constitution invariably fall towards the overcautious extreme; they could be diagnosed as having endogenous depression. Others might always become overconfident, were it not that regular maniacal overconfidence would almost inevitably lead eventually to such misfortune as would predispose to depression instead--so it is not strange that mania without depression is rare or non-existent. Still other individuals would be not particularly biased either way, towards overconfidence or towards overcaution, and so could potentially experience both manic and depressive episodes.

The cause of the alternation
    Let us now consider why some individuals experience both mania and depression, in alternation. Not only do circumstances change but so do the outputs from our genes. Either cause of variation would reasonably be expected to give rise to at least a little fluctuation in the gain of the feedback loop. Even just a small fluctuation would suffice to change the gain from above to below the unity threshold. And in the event of a manic or depressed person experiencing a transitory reduction of the feedback, it seems reasonably certain that an escape from mania would be experienced as a disappointment, biasing towards depression, whereas an escape from depression would be experienced as a success, biasing towards mania. It seems reasonable to suppose that such bipolar individuals would be in the category of those not particularly biased either way.

Other findings about manic-depressive illness
    It has been assumed above that mania and depression are simply excessive confidence and caution respectively. But mania may involve increased movement, talking, pressure of speech, wakefulness, threatening, assaultive, or dominating behaviour, expansiveness of mood and thought, heightened sense of endurance, perceived high energy, self-referential and grandiose, tangential, circumstantial speech, failure to filter trivial stimuli, and inability to maintain focus (2).
    The relationship of these observations to the theory can be understood through further consideration of the nature of confidence and caution. Thus confidence, the state whose essence is a bias towards commission rather omission, could involve not only increased tendency to perceive 'successes' but also a tendency to lowering of thresholds for activation of a wide variety of psychological functions, and this could be embodied in lowered thresholds for firing of a large number of neurons. Thus could result all or most of the features listed above.
    Account must also be taken of the fact that an important feature of depression is sadness, in contrast to the euphoria of mania. The theory faces no peculiar problem at this point, as it is hardly controversial to state that people are made sad by failures and happy by successes.
    This theory accords well with a polygenic-multifactorial causal model with threshold. Constitutional and situational factors would determine the amount of feedback gain applying at a given time, and the threshold would be a specialist-specified frequency/duration of psychosis (i.e. of locking into an extreme). Such a multifactorial model, as has been proposed for schizophrenia (3), is supported by evidence that bipolar disorder represents the extreme of a heritable trait occurring in normal persons (4, 5).
    Also explained is the reason why bipolar disorder does not become "extinct", and why this is a major weak point of the brain. This is because genes producing the feedback are generally advantageous, and are disadvantageous only when, exceptionally, they cause the threshold to be exceeded. Natural selection would favour a level of feedback somewhat below unity gain, such that the disadvantageous exceeding of unity gain would occur only occasionally. Thus humans experience mania or depression occasionally but not invariably.
    There is also a need to account for the growing evidence that relatives of manic-depressives tend to have above-average levels of creativity and achievement (4, 5). This could be because worthwhile creativity requires a combination of openness to unestablished ideas and critical rigorous disposition. Alternating periods of relative confidence and caution would allow alternation between positing of relatively improbable provisional ideas, and their critical reworking and selection. Relatives of manic-depressives would have high levels of the required responsiveness of confidence/caution, but not exceed the threshold with consequent locking into the excesses that are mania and depression.
    Readers may note that this theory invokes a principle already employed in the theory of autism (6), namely that of an advantageous process taken to excess. I do not suggest that this principle applies to all disorders; it should be noted that autism and bipolar disorder share the unusual characteristic of being associated with aboveaverage achievement or socioeconomic status of relatives (regarding which see ref. 6).
    A further feature of the disorder is the prodromal symptoms, which tend to precede and follow psychotic episodes. All these symptoms are in fact characteristic of neurosis (e.g., anxiety). The full explanation of bipolar disorder must await publication of my theory of neurosis.
    It has further been noted that bipolar disorder has some tendency to be associated with schizophrenia, in family pedigrees and in lack of a clear diagnostic boundary. According to a theory of schizophrenia, in preparation, schizophrenia involves another positive feedback loop acting in excess. The link between the two conditions may thus be proposed to be a manifestation of common factors affecting the gain of such feedback loops.

Further suggestions and implications
    It should be noted that the feedback loop is not entirely internal but involves an interaction with external stimuli. Many internal factors, normal or abnormal, could affect the sensitivity of neurons involved in the feedback loop. And aspects of the psycho-environment could also be relevant, for example the degree to which circumstances contain indicators of real success or failure, that is the amount of negative feedback the person receives. This negative feedback corresponds to some extent to the popular concept of feedback as information of results, whether encouraging or otherwise.
    The theory suggests that certain circumstances, where there is a lack of corrective negative feedback, will tend to precipitate bipolar disorder; and convversely adequate negative feedback may help as a corrective or preventative. The theory also suggests that manic-depressive illness would not be confined to humans.
    A two-way causal process remarkably similar to that presented herein has been observed in vervet monkeys (7): rise/fall of status causes increase/decrease of serotonin, and increase/decrease of serotonin causes rise/fall of status (via dominance behaviour). This suggests that status is equivalent to success and that high serotonin is equivalent to confidence. In accordance with this, depression is generally characterised by low serotonin, and antidepressants increase serotonin and sometimes cause mania (8). However, the extent of validity of this parallel remains unclear at present.

References

1. Clarke, R. P. (1994). Draft of a theory of manic-depressive illness and endogenous depression. New Ideas in Psychology 12, 56-60.
2. Carpenter, W.T., & Stephens, J.H. (1980). The diagnosis of mania. In Belmaker, R. H., & van Praag, H. M. (Eds), Mania: An evolving concept. pp. 7-24. New York: Spectrum
3. Gottesman, I.I. & Shields, J. (1982). Schizophrenia: the Epigenetic Puzzle. Cambridge University Press.
4. Coryell, W., Endicott, J., Keller, M., Andreasen, N., Grove, W., Hirschfield , R. M. A. & Scheftner, W. (1989). Bipolar affective disorder and high achievement: A familial association. American Journal of Psychiatry 146, 983-988.
5. Richards, R., Kinney, D.K., Lunde, I., Benet, M. & Merzel, A. P. C. (1988).Creativity in manic-depressives, cyclothymes, their normal relatives, and control subjects. Journal of Abnormal Psychology. 97, 281-288.
6. Clarke, R. P. (1993). A theory of general impairment of gene-expression manifesting as autism. Personality and Individual Differences, 14, 465-482.
7. Brammer, G. L., Raleigh, M. J., McGuire, M. T. (1994). Neurotransmitters and social status. In Ellis, L. (Ed), Social Stratification and Socioeconomic Inequality, Vol 2. pp 75-91. Westport CT, Greenwood.
8. Vieta, E., Bernardo, M. (1992) Antidepressant-induced mania in obsessive-compulsive disorder. American Journal of Psychiatry 149, 1282-3.