Tobacco Documents Online

that this tranWmitter excites neurones in many regions of the brain including! the mcdu11ary and, mesencephalic reticular formation, lateral and medial geniculate nuclei, caudate nucleus, ventrobasal complex of the thalamus, cerebellum, inferior colliculus and! the Betz cells of' the deep pyramidal layer of the cerebral cortex (Phillis, 1'970)1. The cortical cells and caudate nucleus had' definite muscarinic! receptors and were relatively insensitive to nicotine while acetylcholine receptors in the geniculate nuclei, centrobasal thalamus and reticular formation~ nuclei had both nicotinic and muscarinic properties. Cholinergic inhibitory neurones with mixed.nicotinic and muscarinic receptors have been found-at the cortex inn layers II, III and IV of the primary sensorimotor auditory and~visual areas and these seem, to be cholinergic inhibitory interneurones. Im spite of the clear evidence that the: cholinergic neurones at the cortex are predominantly muscarinic, we have seen that "smoking"' doses of nicotine (e:g. 20ug/kg, in the cat) produce! excitation of cortical cells (Knapp, Kawamura, & Domino, 1962; Armitag~e et al, 1969) and release of acetylcholine at the cortex (Armitage et al, 1969). In. the study of Kawamura and Eomino (1969) bloodi pressure was kept.constant with drugs so that the effect of nicotine was not due to this change. Cortical acetylcholiine release and'cortical excitation can be produced by stimulation of the mesencephalic reticular formation and this phenomenon " can be reduced in, one hemisphere by a unilateral destruction, of this region ipsilaterally (Celesia & Jasper, 1966). In a neuropharmacological analysis , of the effects of "smolcing" doses {20ug/Icg) of nicotine after destruction of' the midbrain, (Domino,1967; Kawamura & Domino, 1969) nicotine produced ' cortical desynchronisation and hippocampal synchronization of the EEG in cats with a caudal midbrain transection at the jun;.i+ioni of thepons in 10

exactly the same way as intact animals given nicotine. After bilatcral' lesions in the teginental region of the mid+-brain, nicotine in doses up to five times the 20iug,/kg "smoking" dose did not activate the cortex. Clearly, nicotine's action on the cortex depends on an intact tegmental reg;iony The ventral tegmental region of the mesencephalic reticular formation is the origin of a cholinergic pathway which proj'ects to the cortex (Shute & Lewis,, 1967) and there is good evidence that it excites the pyramidal Betz cells at the sensory cortex and produces electrocortical activation (see review by Warburton, 1981). The most parsimonious conclusion is that "smoking" doses of nicotine ascend in the carotid artery and' excite nicotine receptors on the mid-brain tegmental-neocorticaT cholinergic pathway. it does not act directly on the cortex but the outcome of' activation of this pathway is the release of acetylcholine at the cortex and cortical desynchronisation of EEG of cats. The following sections will critically examine attempts to relate changes iri the electrical activity of intact human brains following cigarette sanokingi ana discuss how the data from animal and human studies may proviidieconvergingi, information about the role of nicotine~ in smoking. 4 SMCKING AND ELECTR00DRTICAL ACTIVITY The appeal.of electrocortical indices is that potentially they can provide us with more direct information about the effects of tobacco smoking on the central nervous system (MIS) of humans. Traditionally, the. EEG has been used as a measure of tonic CNS' "arousal". For example, it is ~ ~ commonly asserted that a shift from high amplitude 8-13Hz to low, aamplitude N' ~ 13-2©Hz activity indicates an increase in alertness. The typical approach ~ i's to compute the level of alpha (8-131iz) activity, the dominant alpha ~ O frequency, the total energy or the variability of EEG activity. It is ~ 11.

assunied~ that the I;EG is a reliable measure of the "etate" of the CNS. However, Gale and Edwards (in press) argue that the EEG may also be a, sensitive measure of phasic events under the appropriate conditions. At this point we must warn the reader that the concept of "arousal" appears throughout the literature as an important explanatory concept. However, this single term is used to interpret datp from a wide variety off studies. Such ubiquitous and often uncritical use of the construct has seriously undermined its explanatory power. The crucial implications for studies of the electrocortica:l co relates of tobacco smoking; are: (i) there is considerable physiological evidence; against the notion of a unitary construct of arousal (e.g. Lacey, 1967); (ii) tYie term has been used in convenient post hoc explanations of unexpected results; (iii) each experiment should be analysed to identify. • important: sources of arousal. Gale (1981) has outlined' an arous l model for the laboratory which has nine components which clearly emphasises the complex nature of arousal-performance interactions. Through systematic identif ication of significant sources'of variability we will be able to generate specific predictions. The alternative is the traditional assortment of "one-shot"' studies providing equivocal results. The averaged'evoked potential and the contingent negative variation (Grlv), collectively, referred to as event-related potentials (ERPs), reflect the response of the- brain; to stimulation. These complex waveforms are external manifestations of CNS processing of simple sensory information and, more important psychologically, complex cognitive processes (Donchin, 1979, 1981). N G N ~ We have described' how n;icotine can cause important changes in the C11 activityof cortical neuro~nesindnimalsa~nd since measurements~ of the ~ ~ 12

electrical activity of the human brain are known to reflect, in part, cortical changes it is reasonable to assume that Ei7G and E1T indices will be systematically related! to the effects of nicotine on the brain. 4.1 SGUECE' OF THE EEG 1y:ND EF2Ps Unfortunately, recording,s from intact humans can only provide gross indices of underlying neuronal activity because they are restricted!to activity picked up by electrodes placed on the surface of the scalp. "Electrodes on the scalp record, mainly the summated! electrical changes of the underlying cortex; they may also record some potential changes• generated in distant parts of the brain and potential-changes produced outside the brain. The amplitude of the recorded potentials depends on the intensity of the electrical source, on- its distance and!spatial orientation, and on the electrical resistance and capaci'tance of thee structures between the source and the recording electrodes. These factors favor the recording of potential changes which (a), occur near the recording electrodes, (b) are generated in, a large area of tissue and (c) rise and fall at slow speed'."' (Spehlmann, 1981; pp 15-16). It is no surprise therefore that the precise sources of human EEG rhythms are by no means completely understood but it is certain that sub-cortical neurones p3,ay a role in EEG activity recorded at the scalip. Consequently, its value ini the analysis of brain, function depends on systematic study of the relationship of changes in the EDG to ongoing behaviour. In the same way, ERPs are statistical measures of' the neural response pattern underlyingi the electrode in which much detail is lost and it will be less thani fully representative of the activity at individual neuroncs/synapses since, for example, some brain processes may not create electrical fields which are measurable at the scalp. 13'

Knowledge of the source of. spontaneous EEG rhythms and! ERPs is important because we need'.to know how different brain areas are involved iin different brain processes. Topographical mapping of EEG' and ERP by multiple site recordings of scalp: activity under a variety`of stimulus conditions will provide information of primary sources of activity.".,...it is incontrovertible that scalp-recorded ERPs are produced by patterns of activity associated' with different neuronal aggregates.... wherever those aggregates are, the sequence with which they are activated and the degree to which they interact with each other reflects intimately the transmissionn of information: and the activation of the information-processing activities withini'untracranial structures.... an ERP component is a subsegment of the ERP whose activity represents a functionally d'istirict neuronal aggregate....A component is a set of potential changes that can be shownn to be functionally related to an, experimental variable or to a combination of experimental variables.... Electrode site is one example." (Donchin, Ritter & MicCallum, 1978; p. 353). 4.2 rIEZ'HODOIi©GIC'AL, ISSUES IN SMOKING RESEARCH Electrocortical measurements associated with smoking present specific challenges to the experimenter. 4.2.1 Sampl ing There must be a theoretical or empirical reason for the time over which the electrical activity is analysed: it would seem logical to link it N O to the time course of nicotine. The EEG~ can be sampled continuously in jy iA milliseconds, seconds, minutes and even hours with the samples partitioned' N CA according to the questions asked. Such flexibility of measurement is a~ ~ source of considerable power only if the rEG is sampled' over the periods ~ 0. 14

when nicotine is known to be active in the brain and also when t})e individual i~s exhibiting any behavioural or experiential effects of smoking. 'Ihis has not been a feature of previous studies. .Similarly,, in ERP studies, the technique of signal • averag,ing depends on repetition of the same stimulus on the assumption that the individual's response to, this event is similar on each occasion. This takes time and since the body concentration of the drug varies across time, reaching a peak and then declining it is even more crucial than in EEG studies that the samples used to make up the ERP are taken during the period when the drug is present and' maximally active. No ERP study of smoking has exercised this. level of care. 4.2:2 Analysis Many different 7nethods have been, used- to quantify electrocortical activity. The wide variety of inethodsused to quantify the EEG - e.g. Fourier 'transf'orms, alpha index, alpha frequency, alpha, abund'ance,, mean dominant frequency - makes it very difficult to compare data derived' from different laboratories and to construct a model of the functional significance of the EEG. A minimum requirement wouldi be to measure a range of frequencies (1-30Hz)'. Mere measurement of gross alp.'la characteristcs (8- 13Hz) may be missing valuable information because other frequencies (betaa and theta, for example) may respond differentially to task conditilons. (Schacter, 1977). ERP analysis also requires a systematic strategy for identifying ERP components (see Donchin, Ritter and McCallum, 1978)1. The crucial point is that electrocortical measures should be carefully integrated with the experimental design and not tagged on as an afterthought. The interested reader is recommended to consult the recent methodological rev~iews~ by dohnso~~n~ (1980; EEG),~ Picton (,11~~98!©; ERP) and; VConnor (19'80; CNV).. 15

4.3' ;,PIOKING 1,TND '111C E3UMAN BRAIN In order to assess: thc psychopharmacological effects of nicotinc, certain, controls must be built into the experimental design: these have generally been ignored in the literature., First, the- state of the subject before smoking has rarely been manipulated in, smoking; research. Thee individual's pre-drug,state must be assessea, since the response! to the drug, may vary as-a function of this initial state., Second, control over drug dbse is vital because many drugs have nonlinear dose-effect relationships. Third, the time course of the drug's activity should be known. These points are particularly apposite in smoking, studies because of the special qualities of nicotine: rapid absorption and wide distribution in the brain together with quick removal. Nicotine is relatively selective by acting on choliinergic neurones but could have simultaneous effects on different cholinergic•brain systems and what is recorded at the scalp might be a function of what demandis are made of the individual's informatiom processing capacities. Here we see a clear case for topographical recording. Furthermore, the relatively brief action of nicotine suggests the drugg can be used as a phasic stimulant which means that the experimenter must take care if he is to synchronise the recordings with this active period,. In an earlier assessment of the EEG effects of nicotine and tobacco smoking Murphree (1974) makes the bold claim that "...since the time of HANS BERGER, certain electroencephalographic (EDG) findings have been known to correlute very closely with behavioral states and' with shifts in those states" (p. 22). Our position is that this is exactly, the opposite: we know relatively little about the relationship •between EL+:`G changes and behaviour andi therefore, if we do not collect information concerning the behavioural [ 16

efiects of drugs concurrently with eiectrocortical measures then we will not be able to identify the functional significance of these changes. Nlurphree also makes the: strange statement that although the drug effect may be the same the EEG changes may be differentl1: "The, point is not ..., that the, in'stial state determines the drug, action as such, but that the subject begins somewhere along the parallel spectra of EEG' and behavioral state and! that the drug, acting as a constraining influence, pushes him: toward other parts of those spectra. The same -kind and degree of drug effects in subjects with different initial states may then produce records having very different appearances." (P. 23). We argue that the point is that unless behaviou;ral andl experiential . data are available in such circumstances there is no way of'verifying that a drug has the same effect on individuals when their brain shows signs of something different happening~ Indeed, one of the most serious criticisms of studies of the effects of' tobacco smoking: on electrophysiological activity is that subjects are maintained at rest and unoccupied prior to and after smoking. For example,. Murphree, Pfeiffer and Price (1967) recorded the EEG of smokers puffing andinhaling,on lighted and uniighted cigarettes in a supine position with eyess closed: hardly normal smoking behaviour. There is little positive evidence for the validity of these contrivedllaboratory experiments and, as a result, there will always be problems with.generalising to active smokers in "real-life" settings. Prior to an evaluati on of the literature, certain necessary design~ constraints will be outlined in order to provide a yardstick for assessing! the quality of the data: obtained. 4.3'..1 Smoking Controls The typical design of these studies compares electrocortical activity 17~

of smokers beforc and after smoking: a study of the acute effects of smoking. While this is a logical strategy, such an apparently simple d!esigm must incorporate appropriate controls. Recording during smoking has b::cn ignored due to the obvious difficulties with movement induced artifacts. However,, with careful control over the smoking behaviour such, studies should be possible and might provide useful informe'-ion about the immediate effects of nicotine. The main difficulty would be that the subject might not be smoking naturally and the signif icance of the datai for normal smoking will be. lessened,, A first step in solving; these problems would be to incorporate a detailed's analysis of' normal smoking behaviour. Without adequate controls for the smoking action any electrocortical changes cannot be confidently attributed to nicotine. The problem of whatt constitutes an adequate control is a matter of debate and different- researchers appear to havie their own favourite. At the very least a nonsmoking condition is needed to control for the effects of the laboratory environment. By including the smoking of a nicotine free cigarette in the design it should be possible to partial out the effects of the environment, smoking action and nicotine on brain activity. Suitable controls are more readily available in nicotine tablet or injection studies but, as we have argued above, there are difficulties in extrapolating from the results of these studies to the results of cigarette smoking; particularly when the dose administered bears little relation to that typically obtained, fr= cigarette smoking (see Section 4. In addition to~analysing the "cigarette end" of'smoking, useful information can: be gathered about subjective aspects of' smoking~ behaviour.. Individual differences in; smoking, behaviour and smoking motivation have been~ negleftedin the literature with the exception of two CNV stud~ies 18

(Ashton, h3illman, Telford &'i:hompson, 1974; Binnie & Comer, 197'3'). 4.3.2' The Cigarette Only Knott and' Venables (19 7', 1979) and! Binnie and'. Comer (197'a), provide machine smoking estimates of the nicotine and tar deliveries of'the test cigarettes. 2'be remainder merely report that subjects smoked one or two cigarettes during the experiment. While the Fhl-Acal characteristics of the cigarette only provide limited information about the dose of nicotine obtained,during smokingi (see next section) these data can assist comparison across studies.= 4.3.3 Smoking Behaviour "Cigarette smoking .... is a drug-delivery system for nicotine, "tar"' and carbon monoxide that affords no .... straightforward' way of monitoring dose. Although the words on cigarette packs and advertisements appear t© indicate. the dose of tar and nicotine to be found' in a cigarette, smokers can easily double or triple yields beyond the nominal levels by taking,more frequent, larger, or high velocity puffs.." (Kozlowski,, 1981;p213'). Standard smoking machine deliveries do not ariequately reflect normal human smoking: because the machines take a 2 second, 35 ml puff on, a cigarette per minute until a fixed length butt is obtained: human smoking is much too variable to b properly represented by such an arbitrary average puffing schedule. Thus, even if studies of the electrocortical effects. of smoking had included standardl machine delivery information we would still not know- the nicotine dose that the subject obtained from snriking that particular cigarette. A number of parameters of smoking bchaviour are measurable: number of puffs per cigarette, puff duration, inter-puff' interval, puff volume, butt length,, butt nicotine analysis,- percentage tobacco burned, 19