OMAN 297 e orientation of the vehicle is described in engineering t gular and linear position, velocity, acceleration, and often other factors. Engineers are tly confronted with not directly measured at all. It is a complex multi-input, multi- nown 'observer(as shown in Fig. 1) of actual vehicle orientation state in the traditional way is not FG.1. oi vehicle onen. isy and do not directly mesetor quantities eed to be controlled. Vehicle must follow op param- orentation command signals in spite of exogenous disturbing forces. the entire system. In control engineering parlance, st systems are formally called"output feedback"optimal control is the way in The observer contain of cl these models to calculate what the actual feedback ser easurements should etween the expected and the actual feedback measurements is bserver vehicle orientation state towards reality, i.e., the F.2. the predicted measurement will fade to the level of the parallels Fig. I ontrol based on observer theory (Oman 1982). Note that the structure odels is correct. In practice, a sustained high level of error in and linear of all the parts of the body and higher are the dif- rential equations that are the observer intemal dynamic ferential bing the body and sense organ s m us n gy betwen the"expected feedback manas, Me neads basedon", the estimated onend. The CNS must infer the bodys orientation and control body tion "state is used by the CNS sense organ model to comp movement using a limited set of noisy sensory signals. Fror natural behavior of the body, i.e., have an"in st, is familia the perspective of control engineering, the problem is familia ector nearly cancels polysensory afference. when it does the difference (the*ser mal model of the model predictions towards reality and indin he dynamics of the body and maintain a continuous estim spatial orientation of a traction of an ung sensory inputs she parts. Based on observer critical functional role for conflict signals in daily life. Clear f the sp ld be evaluated by sub- we need not postulate that sensory conflict signals exist just to In normal life, we experience exogenous forces only briefly analogy, the (e. g, when we stumble over an obstacle), so increas actual orientation"'vector might here consist of the angular sory conflict are normally only temporary. Sustained sensory
CAN, I PHYSIOL PHARMACOL onflict would be expected under conditions of sustained exog- 回时四回 istics of the sensory organs or the body itself were sudde 978; Held 1961). People experience such conditions when, conflict indi- difa <hat the CNS needs ta revise the observer intemal model sumably the CNS monitors the average level of sensory con- fict. and whe high, the dynamic model equations are ed so that average is reduced. This " model flict signals in daily lif Presumably, the CNS observer them in 3. Mathematical formulation of model shown in Fig. 2(Or Exactly how an observer sensory conflict vector might be us- matrices A. B, and S corespond to the"neun ed to steer intemal model predictions is described briefly (1978)qualitative model. The sensory conflict below and in Fig. 3. Readers who are familiar with matrix subtracting actual sensory input a from expec ory input algebra will recognize that the large number of linearized dif- See text for additional details bodys neuromuscular and sensory systeing state variabe tt gse that experienced drivers and pilots consider their vehicles ta. when running, jumping, dancing) do not cause increased sen- equations. Additional available in Oman ( ppendix 1). The equations hay the t extensions of their own bodies by using an appropriately 塞=Ax+Bu anticipate the effects of their own control commands on ntly 3]u=田+0。 flict. In contrast the sa orces by the passengers, who nts of the state differ for that the abilit xplain it, and am加FMmm了m油 essential reason to distinguish between body and sense exist in the CNS, at least for certain sensory modalities sented by an analogous state differential equation shown in the signals comesponding to the efference copy vector a. The ne iom half of the figure. CNS observer estimates are denoted efference copy effect might in fact be distributed betw of Reason's(1978)more qualitative model. The sensory con-(See Oman by itself, exactly cancels We shall om expected sensory input S mally identified later(Conclusions), after considering the nature of y orginates fr tion cue inputs ne and sen- the emetic linkage ry noise n,. The confl lculated using an optimization technique defined by Kalman a and Bucy (1961), which lightly weights noisy modalities Experiments on the dynamic character of symptoms state,the estimated state vector is driven t of the estimated iscussing models for the emetic linkage resp state, and the component of the conflict vector magnitude due and Oman 198 the rate of change of the estimated state via the 986: Oman 1987: Rague and Oman 1987: Eagon m Bm. Because of this term, active body movements 1988). One purpose of these studies was to define the dynamic
haracter of the emetic linkage and response pathways. We written questionnaire or assembling structures on a tabletop types of stimu ng out-of- using soda cans while wearing prism goggles was provocative chair, and head movements while walking A revised model for symptom dynamics The authors (1982) motion sickness model ial flict and overall discomfort and nausea in motion si sickness lev the Pensacola diagnostic the cument version is shown in Figs. 4 and 5. ant c index methodology( Graybiel et al. 1968) cy of response of our i from Fig. 3. Because of the band cts vanes Is re imposed on signals involved in orientation percof the ely that the componen ed temporally with the appearance of stomach discom- weighted and added together correspond to Reason's(1978) he result is referred to as the ort or nausea. (Some of these consistently show a transient"neural mismatch sign 10 min. When the stimulus is first removed, nausea fades However, stimuli that presumably produce sensory conflicts of to elicit a moderate pallor response, skin color retums to weighting the various confi (i. e, ipate in experiments last- weighted relatively heavily in the since people withor g several hours. Nauseogenic stimuli are ly find that when tho allow symptoms to subside, wi provocative stimulus is reapplied may thus exert their major sick-making effects indirectly re rapid. After the stimul Sooner ad skin removed, and nausea produced in the heavily weighted aded away, usually provide a reliable cue that the subje n- visual and proprioceptive conflicts under prism goggle sensory writing derable additional time to recover completely. It is as if a block building, absent any comitant head motion or ve bular conflict, th ninutes to fully develop, but once this shift is established, thy weighting factors should ably many other concurrent physic zero, The implication rably more sensitive to subse subliminal visual and proprio allor, nausea, and overall discomfort tive emetic coupling. Even if the vestibular sensory conflict port waves of signal paths in these people are inactive, their functional Then the sensitized sta reached,many subjects find it difficult to pace themselves pathways above subliminal levels owing to the lack of the ther types of nauseogenic stimuli which previously seemed vestibular contribution 5. the neural mismatch information acts via two parallel, interacting pathways(c and had been standing before, stumbled on a curb and, nitude estimate mod址 Magnitude estimates ar 1957 d"‘slo resold- power law element and resulting in a nausea mag- reported that the disparity between the seen and audible posi- with an exponent of about two. Susceptibility to motion sick king produced twinges of nausea. A third gog. ness is determined in the model not only by the amount of sen- gles subject who had been resting for several minutes while his sory conflict produced, but also by the fast comfort decreased was then asked to draw the face of lock on a blackboard and was so disturbed by the disparity a generalized adaptation from one different nauseogenic stimu- d be vomited with little waming. Almost all of the symptomatic output pathway subjects in Eagon's(1988)study reported that filling out a The parallel arrangement of the fast and slow pathways and