Models of memory dysfunction? A comparison of the ... - Springer Link

Psychopharmacology (1991) 103: 83 90

Psychopharmacology © Springer-Verlag 1991

Models of memory dysfunction ? A comparison of the effects of scopolamine and lorazepam on memory, psychomotor performance and mood H. Valerie Curran, Fabrizio Schifano, and Malcolm Lader Department of Psychiatry, Institute of Psychiatry, De Crespigny Park, London SE5 8AF, UK Received January 30, 1990 / Final version June 15, 1990

Abstract. The effects on memory, psychomotor functions and mood of intramuscular scopolamine (0.3 rag, 0.6 rag) were compared with those of oral lorazepam (2 mg) and placebo. Thirty-six volunteers took part in a doubleblind, independent groups design. Subjects completed a battery of tests 1 and 3 h after drug administration. Both doses of scopolamine produced levels of sedation comparable to that produced by lorazepam. The time course of effects of scopolamine and lorazepam differed but the pattern of psychomotor impairments and amnestic effects produced was very similar. In terms of mood, lorazepam had an anxiolytic effect whereas scopolamine increased ratings of anxiety. Levels of sedation, indexed by either subjective ratings or motor retardation (tapping speed), were related more to psychomotor performance than to performance on memory tasks. The results suggest that benzodiazepines and scopolamine have similar amnestic and sedative effects and as such may not offer distinct models of memory dysfunction. Key words: Scopolamine - Lorazepam - Memory - Sedation - Psychomotor performance

In a widely cited paper, Weingartner (1985) suggests that different forms of human memory dysfunctions can be modelled by the administration of different centrally acting drugs to normal, healthy subjects. He argues that whereas scopolamine (hyoscine) produces transient amnestic effects in normal subjects similar to the deficits observed in patients with Alzheimer's disease (AD), benzodiazepines (BZs) produce effects more similar to the anterograde amnesia typical of patients with Korsakoff's disease (KD). Based upon the "cholinergic hypothesis" of AD, there have been many studies of the amnestic effects of the anticholinergic drug, scopolamine, on normal subOffprint requests to: H.V. Curran

jects (see CoUerton 1986; Sahakian 1987 for reviews). Similarly, there is a broad literature on the memory impairments produced by BZs (see Curran 1986, for a review). One problem with using scopolamine to "model" the cognitive deficits of AD stems from the question of how specific are the drug's effects on memory. Although it is clear from a number of studies that scopolamine impairs the performance of normal subjects on tasks involving memory, it also impairs performance on tasks which have no obvious memory component - for example, finger tapping speeds (Nuotto 1983) and simple reaction times (Callaway 1984). In attempting to support the cholinergic hypothesis, many studies have focussed only on the effects of scopolamine on memory and excluded other assessments. Other studies have demonstrated impairments of performance on "attentional" tasks (e.g. Dunne and Hartley 1985; Broks et al. 1988), and it has been suggested that memory impairments are secondary to scopolamine's effects on attention (e.g. Sahakian 1987). More recently, evidence has been presented which implies that a main locus of scopolamine's effects is on the central executive component of working memory (Rusted 1988; Rusted and Warburton 1988). Given that scopolamine has been found to impair a range of executive, attentional and memory functions, as well as performance on simple motor tasks, it becomes relevant to ask whether scopolamine has any specific cognitive effects which are not secondary to its sedative properties. The sedation produced by scopolamine has been noted many times, but little attempt has been made to relate this to impairments in performance on attentional or memory tasks. At least in part, this reflects the problems of dealing with an ill-defined concept (sedation, arousal, alertness, activation) which is also empirically problematic in that the various physiological, psychomotor and subjective measures do not intercorrelate well (Eysenck 1982). At the same time, the problem remains that if a drug is sedating a person, there will be global impairments across a range of tasks and this may preclude conclusions about specific cognitive effects.

84

That scopolamine's amnesic effects may be, in part at least, secondary to its global sedative effects is also implicated by the few studies comparing scopolamine with a benzodiazepine. Benzodiazepines (BZs) also have sedative (as welt as amnestic) effects and are prescribed as "sleeping pills" as well as anxiolytics. It further appears that some BZs, like lorazepam, are more potent amnestic agents than others, such as oxazepam (Curran et al. 1987, 1988). Studies comparing scopolamine directly with diazepam or nitrazepam have not found consistent differences in their effects on memory (Ghoneim and Mewaldt 1975, 1977; Jones et al. 1979; Frith et al. 1984; Richardson et al. 1984). However, surprisingly, none of these studies have assessed the covariation between the drugs' effects on memory and their sedative effects. To support Weingartner's suggestion that scopolamine and benzodiazepines may offer two different models of memory dysfunction, one would need to demonstrate that these drugs had differing effects on memory functions reflecting differences in clinical syndromes of cognitive impairment. Further, given that patients with either AD or KD do not suffer from chronic drowsiness, one would need to differentiate the sedative effects of both BZs and scopolamine from their amnestic effects. It was of interest, therefore, to compare the effects of these drugs directly on a wide range of tasks and to relate arousal changes to performance changes. With this aim, the present study compared lorazepam with two doses of scopolamine and a placebo. A battery of tests was used to give a broad assessment of the memory and nonmemory effects of the drugs. The battery included measures of working memory, secondary memory and psychomotor function. Psychophysiological measures (of peripheral anticholinergic effects and central "arousal") and subjective ratings were also employed.

Test battery This consisted of tests of working memory and secondary memory, psychomotor function, psychophysiological measures and subjective self-ratings. Order of the three test versions was counterbalanced across subjects and design. A fourth version of the tests was used in the training session.

Working memory tasks Digit span. Stimuli comprised digits 1-9 in random sequences presented auditorily at the rate of one digit per second. The subject's digit span is one less than the sequence length he fails to repeat on two successive trials. The task was used as assessment of verbal working memory.

Mental rotation. Versions of this task were based on the task developed by Shepherd and Metzler (1971) and used in a study of scopolamine by Rusted (1988). It was used as an assessment of visuospatial working memory. The subject has to determine whether a capital letter, presented on a VDU screen, is correct or mirror reversed. The letters R, J and F formed the three main versions of this task. The letters were presented at one of eight different orientations from the vertical axis, with five correct and five mirror-reversed letters for each orientation. The subject responded by pressing a YES or NO key, and all responses and reaction times over the 80 trials were recorded automatically. Baddeley reasoning test. This task is a computerised version of that described by Baddeley (1968) and used in a study of scopolamine by Rusted and Warburton (1988). It loads onto central executive function. Subjects were presented with a series of statements on a V D U such as "A precedes B .... BA"; "B is preceded by A .... AB"; "A follows B .... BA". They were asked to indicate whether the statement described the correct order of the adjacent AB letter pair by pressing either a "true" or a "false" response key. Sixty-four trials were presented involving four different grammatical constructions and responses and reaction times were recorded automatically.

Secondary memory tasks Materials and methods

Subjects Thirty-six healthy volunteers (24 women, 12 men) aged between 21 and 49 years (mean: 27 years) participated in the study. All subjects gave written informed consent.

Deson An independent groups design was used whereby subjects were allocated to one of four treatments: lorazepam (2 mg), scopolamine (0.3 mg), scopolamine (0.6 rag) or placebo. Allocation of subjects to treatments was random apart from balancing by sex (six women, three men in each group). Double-blind procedures were followed using a double-dummy technique such that each subject received one capsule (lorazepam 2 mg or lactose placebo) and one intramuscular injection (scopolamine or saline placebo). Subjects were not allowed alcohol or other CNS drugs from 24 h before the testing day to 12 h after testing had finished. Four to seven days before they were to be tested, each subject was trained on all the tasks they were to do in the study. On the testing day subjects were tested in the morning before drug administration and again 1 and 3 h after the drug. Each test session took 50-60 min.

Paired-associate interference task. This task was based on an A-B, A - C interference paradigm used by Mayes et al. (1987). It has four short stages. First, subjects are asked to read aloud and try to remember a series of 16 word pairs (A-Bs) that are highly associated (e.g. towel-bath, letter-envelope, coal-black). ARer a filled interval, they are shown the first word of each pair and asked to recall the word that went with it. They are then immediately shown a second set of word-pairs to read aloud and remember. This second set (A-Cs) has the same first word (i.e. A-word) as the first set but a different, highly associated second word (e.g. towel-rail, letter-post, coal-miner). After an identical filled interval, they are then given the first word from each pair and asked to recall the word that went with it the second time. Interference is assessed at this point as the number of B-words recalled instead of C-words. Word pairs were presented for 4 s each. The two intervals were "filled" using two 50-s trials of a pursuit rotor task in which the subject tries to follow the track of a rotating light source with a light-sensitive probe (Forth Instruments Ltd). Time in contact with the light source was recorded over five 10-s sample periods for each interval. This task was used to give a measure of procedural learning.

Prose recall. In this test, the subject is asked for written recall of a "news bulletin" type of story presented via a tape recorder (i) immediately after hearing it and (ii) after a filled delay of about I5 min. Versions used were from the Rivermead Behavioural Memory Test (Wilson et al. 1985). Each story consisted of 21 "idea units"

85 and between 60 and 65 words. Scoring was 2 points for perfect recall, or an exact synonym of each idea unit and 1 point for partial recall or partial synonym.

Word fluency. This task requires timed retrieval from semantic memory. Subjects were asked to name out loud as many (1) instances of a given category (2) words beginning with a given letter of the alphabet as they could think of. Ninety seconds were allowed for each part of the task. All responses were tape-recorded and scored in six blocks of 15-s intervals. The categories used were four-footed animals, countries and colours, and were selected from the Battig and Montague (1969) norms to approximately equate numbers of high frequency instances. The letters used were B, F, and M, and were selected from the Oxford pocket dictionary to approximately equate the numbers of common words each letter began. In the practice session, the category used was vegetables, and the letter was C.

random digit sequences is recorded with errors. The symbol copying test (SCT) was used as an index of manual writing speed over a 90-s interval. Tapping rate was used as an index of motor sedation after File and Lister (1983) (the subject presses a key as quickly as possible for 60 s). The digit symbol substitution test (DSST), a 90-s recording task from the Weschler Adult Intelligence Scale (Wechsler 1955) was also used.

Physiolooical measures Peripheral anticholinergic effects were monitored using the salivary flow measure described by Peck (1959) and the visual near point measure described by Kopelman and Corn (1988). Critical flicker fusion threshold (CFFT) was assessed using a fixed-pupil technique.

Subjective ratinos Psychomotor tasks Focused attention was assessed using a simple digit cancellation task in which the time taken to score out the number 4 from

Levels of anxiety were self-rated on the State Anxiety Inventory (Spielberger et al. 1969). A visual analogue mood rating scale (Bond and Lader 1974) was used to assess subjective feelings of alertness, calmness and contentedness.

Table 1. Group means pre-drug (0), 1 h and 3 h post-drug administration Placebo 0 Baddeley reasoning task (RT's) (s) Mental rotation: R T normal stimulus R T mirror reversed stimulus (ms)

Lorazepam 2 mg lh

3.01

3h 2.80

0 2.52

lh 3.52

3h 4.49

4~21

Scopolamine 0.3 mg

Scopolamine 0.6 mg

0

0

lh 3.72

3h 3.60

3.39

lh 3.40

3h 3.44

3.46

908

807

683

754

875

784

999

1017

833

t059

t093

1020

1056

835

781

1080

1090

1151

1215

1143

1015

1356

1147

1256

Digit span Paired associate task: A-B A-C Intrusions

7.7

8.2

7.7

8.6

7.9

7.8

7.2

7.8

7.8

8.1

6.8

7.4

13.9 13.6 1.3

12.5 13.0 1.2

11.7 13.2 1.3

14.2 13.7 1.3

11.2 10.6 2.2

9.7 8.6 2.3

14.9 12.4 1.9

tl.3 12.0 2.1

tl.9 11.2 2.8

13.7 12.8 1.6

10.0 9.8 2.1

12.0 12.3 1.4

Prose recall: immediate delayed

28.0 27.4

28.6 25.7

26.6 25.0

26.6 24.9

24.0 21.1

17.6 8.7

27.3 25.8

19.7 16.6

21.4 18.3

29.3 28.0

17.7 9.6

18.3 16.3

73.0

73.6

71.7

77.3

67.1

63.7

66.9

60.6

65.4

75.3

61.7

66.7

SCT

DSST

178.0

183.7

182.3

178.4

160.6

156.6

164.4

153.4

158.0

180.1

158.6

167.3

Intertap interval (ms)

t74.4

172.7

181.1

165.7

175.6

t81.1

167.2

185.2

180.2

167.4

198.5

t85.3

Cancellation (s)

42.5

40.4

40.2

43.2

47.8

52.7

56.0

55.4

54.9

40.6

49.0

47.2

13.5 17.8

18.3 21.6

18.9 20.5

13.9 18.0

15.5 16.0

13.9 13.7

16.4 17.2

13.7 15.8

16.0 18.7

17.2 22.5

9.5 10.4

13.0 15.9

22.2 2t.6

22.1 20.9

24.7 22.8

24.1 22.9

24.3 20.7

23.9 21.8

23.3 21.3

23.3 21.3

25.3 21.1

23.2 22.2

22.0 19.2

21.6 19.2

Pursuit rotor: trial 1 trial 2 Word fluency: letter category Salivary flow Visual near point (cm)

1.42

1.36

1.24

0.83

0.83

0.76

0.99

0.29

0.55

1.10

0.29

0.74

9.5

10.7

10.8

12.1

13.0

13.2

8.2

11.0

10.9

9.9

21.9

C F F T (Hz)

37.4

36.3

36.5

38.0

36.5

36.2

37.7

36.2

36.1

38.0

35.8

24.2 36.1

State anxiety Subjective sedation (MF1) (mm) Subjective excitedness (MF3) (mm)

30.7

30.7

30.7

34.1

33.8

33.0

32.0

34.1

32.8

31.6

40.3

39.7

38.8

47.0

44.9

24.4

48.2

56.6

29.3

61.6

48.7

32.7

74.3

57.4

18.0

19.2

20.2

28.0

16.9

15.4

29.6

26.1

23.8

24.0

31.7

25.8

86

Statistical analysis Preliminaryanalysisof variance showedsignificantpre-drug group differenceson some variables. All data were therefore subjected to analysis of covariance (ANCOVA) wherebypre-drug scores were covaried out from both post-drugscores. Planneda priori contrasts were computed to compare active treatments with placebo; iorazepam with each dose of scopolamine; the two doses of scopolamine. As described below, all active treatmentsproduced marked sedation and so a further ANCOVA was carried out to examinethe relation between indices of sedation and performance on the cognitive and psychomotor tasks. All graphs are plotted as change scores from pre-drug values and group means are presented in Table 1. Results

0 I1)

I

-10 O

o -20 3 hour

Fig. l. Change scores from pre-drug levels on delayed recall of prose for subjects given placebo (o), scopolamine 0.3 mg (n), 0.6 mg (I) and lorazepam (A)

Working memory tasks Digit span. The only significant difference found on this task was between the two doses of scopolamine whereby 0.6 mg slightly reduced digit span at the 1-h testing time compared with 0.3 mg [F(1,31)= 6.1, P