Clinical Neuropharmacology

Department of Psychology, West Valley College, Saratoga, California; *Department of

Psychology, Stanford University, Palo Alto, California; and tLangley Porter Institute,

Summary: Previous studies have not demonstrated a consistent relationship between precursors to acetylcholine (ACh) and memory function in normal human subjects. This experiment (N = 80, college students) employed a double-blind mixed design to test the effect of phosphatidy1choline (PCh) on explicit memory. Dose of placebo and PCh was compared at two levels (10 and 25 g) as was time of testing postingestion (60 and 90 min). With 25 g of PCh, which supplies 3.75 g of choline, significant improvement in explicit memory, as measured by a serial learning task, was observed at 90 min postingestion and slight improvement was observed at 60 min postingestion. Further analyses indicated that this improvement may have been due to the responses of slow learners. This is the first study to test the relationship between a single dose of PCh and explicit memory on normal human subjects.

Key Words: Phosphatidylcholine-Lecithin-Acetylcholine precursor—Memory, explicit—Learning, serial.

Over the last three decades the hypothesis that memory is modulated through cholinergic systems has been studied using a variety of pharmacological strategies on different subject populations. These strategies include the use of acetylcholine (ACh) precursors, anticholinesterase agents, central muscarinic agonists, and central muscarinic antagonists. In lower animal research, it has been demonstrated that drugs affecting cholinergic mechanisms have an influence on memory (1,2). However, there has not been a breakthrough in the attempt to correct memory deficits by manipulating the cholinergic system in patients with Alzheimer's disease (3,4) even though the muscarinic receptor binding sites are not significantly depleted (5). Neither has the role of ACh as a central mediator of memory in normal human subjects been clarified. Pioneering research by Sitaram et al. (6) reported that the ACh precursor choline chloride improved serial learning, while Davis et al. (7) and Mohs et al. (8) found no significant relationship between the ingestion of choline chloride and tests of digit span, memory retrieval, and memory storage. Sorgatz (9), investigating the cumulative effects of fluid lecithin, found significant improvement in word recognition and retrieval by category tasks, while Harris (10), using a single dose of lecithin, found no significant change in several subcomponents of memory. Gillin (11), administering lecithin at three different dose levels, found no significant improvement in either serial learning or paired associate word tests.

In contrast to the inconsistent findings with ACh precursors, increased memory performance has been reported on normal human subjects with the central muscarinic agonists arecholine (6); a slight trend toward memory improvement has been found with the anticholinesterase agent physostigmine (12,13); and memory impairment has been noted with the central muscarinic antagonist scopolamine (6,14,15). Because other pharmacological strategies of modulating ACh at the synapse have produced positive findings and brain concentrations of ACh may increase following the administration of ACh precursors (16-19), we hypothesized that there would be improvement in an explicit memory task as a function of the dose of the ACh precursor phosphatidylcholine (PCh) and time of testing postingestion. Explicit memory refers to the learning, retention, and retrieval of information held in memory for periods >~30 s (14).

Ten college students (4 males, mean age 19.8; 6 females, mean age 19.4) participated in the preliminary study, and 80 college students (40 males, mean age 20.5 years; 40 females, mean age 20.4 years) participated in the experiment after informed consent was obtained. Subjects were screened for milk allergies and dyslexia. Subjects received extra credit points in general psychology for participating in this research. Three subjects dropped out of the experiment before completing the procedure, and four subjects were deleted from the analyses because they did not reach learning criteria (two on placebo and two on TwinLabs PC-55).

TwinLabs PC-55 (TL PC-55) (Twin Laboratories), a highly concentrated, specially fractionated, de-oiled soybean lecithin with increased phosphatidylcholine content was administered to test the effect of PCh on explicit memory. The chemical composition of TL PC-55 was as follows: phosphatidylcholine (55.0%), phosphatidylethanolamine (30.0%), phosphatidyl inositol (3.0%), other phosphatides (8.0%), and tryglycerides neutral oil (4.0%). In TL PC-55, 1 g of PCh supplies 0. 15 g of choline.

Two solutions were used in the preliminary study and four solutions in the experiment. A comparison of dose of PCh, choline supplied, and amount of the TL PC-55 preparation is presented in Table 1. In the placebo mixtures, the gram equivalent of cornmeal was substituted for the TL PC-55 described in Table 1. Common to each 10 g PCh, 25 g PCh, or placebo preparation was 150 ml nonfat milk, 250 ml distilled water, and 75 mg cinnamon. In the preliminary study, 25 g of PCh and the placebo equivalent were administered with 75 mg of curry used for flavor instead of cinnamon. Subjects had free access to distilled water during consumption.

The preliminary study was a double-blind counterbalanced within-subjects design. Subjects were observed at one dose-time level: 25 g of PCh with testing at 90 min postingestion. To determine if this was the optimal dose-time level for PCh effectiveness in an explicit memory task, a double-blind mixed design was conducted. The subjects, 40 men and 40 women, were randomly assigned to one of the four groups: PCh10g60m, PchlOg90m, PCh25g60m, PCh25g90m. An equal number of men and women were assigned to each group. Placebo and PCh were administered to each subject, counterbalancing for order within a group and between the sexes.

The procedure, used in both the preliminary study and the experiment, was conducted on two days. On day 1, subjects ingested either placebo or PCh at 7:30 a.m. after fasting for 8 h. Either 60 or 90 min postingestion, subjects were administered a serial learning task using the nonsense syllables from either list 1 or 2, counterbalanced for order within each condition (placebo or PCh). On day 2, 48 h later, the same procedure was followed. The order of placebo and PCh administration was counterbalanced for each group.

The dependent variable, number of trials to criterion on a serial learning task, was used to assess differences across the four experimental groups. A repeated measures multivariate analysis of variance (MANOVA) was computed with dose (10 or 25 g PCh) and time of testing postingestion (60 or 90 min) as the between subject factors and pharmacological agent (placebo and PCh) as the within-subject factor. The Geisser and Greenhouse (21) correction of the degrees of freedom of the F ratios for the within-subject factor and its interaction terms was performed. Differences in serial learning for the four experimental groups were also assessed by an analysis of variance (ANOVA) and an analysis of covariance (ANCOVA) computed using difference scores as the dependent variable, number of trials to criterion for placebo minus number of trials to criterion for PCh (P-PCh). Two tailed paired t tests were used for within-group comparisons of placebo versus PCh for the preliminary study and the four experimental groups. Means ± SD are reported for each group. The statistical significance criterion for these multiple analyses was adjusted to p = 0.008, according to Bonferroni's approach. Individual differences in learning rates (slow vs. fast learners) were assessed by a Mann-Whitney U test, with z scores corrected for ties, on the difference scores P-PCh.

The main effect of the MANOVA, which combined the data at 10 and 25 g for both placebo and PCh and did not constitute a test of the hypothesis, was insignificant (F = 0. 13; df = 1,77; p > 0.72). The critical finding was a significant interaction between the pharmacological agents (placebo and PCh) and dose (F = 7.71; df = 1,77; p < 0.0069). There was no significant interaction between the pharmacological agents (placebo and PCh) and time (F = 0.21; df = 1,77; p > 0.65).

Further analyses were computed using the difference scores, P-PCh, as the dependent variable. A 2 X 2 ANOVA was computed to assess whether there was a significant interaction between dose and time using difference scores as the dependent variable. The mean P-PCh values for the four experimental groups differed significantly for dose (F = 7.90; df = 1,76; p < 0.006). No significant difference was obtained for time (F = 0.21; df = 1,76; p > 0.65), nor was there a significant interaction between dose and time (F = 2.84; df = 1,76; p > 0. 10). To control for individual differences in number of trials to criterion on placebo, a 2 x 2 ANCOVA was also computed. With the placebo as the covariate and P-PCh as the dependent variable, significance was obtained on the adjusted P-PCh means for dose (F = 8.50; df = 1,75; p < 0.005), but not for time (F = 0.45; df = 1,75; p > 0.51) or dose-time interaction (F = 1.61; df = 1,75; p > 0.21).

Paired t-tests computed for number of trials to criterion for placebo versus PCh for the preliminary study and for each experimental group are presented in Table 2. The data from the preliminary study and the experimental group, PCh25g90m, were combined. A significant difference in mean trials to criterion for placebo versus PCh was only obtained for the combined PCh25g90m group. These results indicate that improvement in explicit memory with 25 g of PCh occurred at both 60 and 90 min and obtained significance at 90 min. Ten grams of PCh did not significantly affect explicit memory.

Comparison of mean trials to criterion for 10 and 25 g of PCh at 60 and 90 min postingestion

Prelim study PCh25g9Orn	Placebo PCh	5.3 ± 2.7 3.9 ± 1.8	9	2.14	0.061
Experiment PCh25g9Om	Placebo PCh	6.1 ± 2.3 4.8 ± 1.2	19	2.58	0.018
Combined PCh25g9Om	Placebo PCh	5.8 ± 2.3 4.5 ± 1.5	29	3.39	<0.002^a
Experiment PCh25g6Om	Placebo PCh	4.7 ± 2.2 4.5 ± 1.9	19	0.60	0.555
PCh10g90m	Placebo PCh	5.0 ± 2.8 5.9 ± 2.3	19	-1.77	0.092
PCh10g60m	Placebo PCh	4.8 ± 2.1 5.1 ± 2.0	19	-0.57	0.573

Paired two-tailed t tests with statistical significance criterion adjusted to p = 0.008, according to Bonferroni's approach.

An ANOVA for number of trials to criterion under placebo administration for the four experimental groups was insignificant (F = 1.39; df = 3,76; p > 0.25), which makes between group comparisons valid.

Individual differences in learning rates for the combined groups at PCh25g90m were assessed by a median split based on a subject's total learning score: trials to criterion on placebo plus trials to criterion on PCh. Because the SD for the slow learners (2.4) was nearly twice as large as that for the fast learners (1.4), a MannWhitney U test, with z corrected for ties, was computed and obtained significance (2.3 ± 2.4 vs. 0.4 ± 1.4; z = 2.19; p < 0.029). This finding supports the possible interpretation that the results obtained for PCh and explicit memory for the combined PCh25g90m group were due to the responses of slow learners.

At the end of each experiment, subjects were asked to rate gastric symptoms (nausea, stomach awareness, and salivation). Each symptom category was rated on a 1 to 5 scale, with a total of 15 points possible. To determine whether the symptoms reported under PCh administration may have affected performance, a correlation was computed between gastric symptoms reported for PCh and a subject's P-PCh score. No significant correlation was found (r = - 0.04; df = 65; p > 0.05). A similar correlation was computed for gastric symptoms reported for placebo and a subject's P-PCh score. No significant correlation was found (r = 0.002; df = 65; p > 0.05). Previous research reporting gastric symptoms with lecithin administration has used higher doses than the ones used in the present study (11).

The main treatment effect reported here represents the first time a single dose of the precursor PCh has been tested and shown to enhance explicit memory in normal human subjects. With 25 g of PCh, which supplies 3.75 g of choline, significant improvement in explicit memory as measured by a serial learning task was observed at 90 min postingestion, and slight improvement was observed at 60 min postingestion. Further analyses may indicate that this improvement in explicit memory was due to the responses of slow learners, confirming a trend observed by Sitaram et al. (6) with arecholine and choline chloride, by Mohs et al. (8) with choline chloride, and by both Gillin et al. (11) and Sorgatz (9) with lecithin.

Our data suggest that the lack of control for subjects' learning rate may have contributed to the inconsistent findings in this area of research. It may be that PCh supplements do not uniformly enhance memory, which implies that the dose-time parameters required to obtain a treatment effect may be contingent on subject variables. Slow learners may have subnormal endogenous levels of choline, which may be elevated by PCh supplements and produce a measurable improvement in explicit memory.

Important to the argument that cholinergic mechanisms have a specific influence on explicit memory is the fact that scopolamine affects tests of explicit memory but does not affect tests of remote, semantic, and implicit memory (14). According to Gabrieli (22), explicit memory appears to have different neural substrates than those that mediate implicit memory. However, since a high concentration of cholinergic fibers have been reported in the visual cortex (23,24) as well as an abundance of muscarinic cholinergic M2 receptors (25), future clinical trials might examine both verbal and visual explicit memory with PCh supplements on subject populations suspected to have subnormal endogenous levels of choline.

The only study reporting memory enhancement with lecithin on normal human subjects was conducted by Sorgatz (9) in 1987. Sorgatz tested the memory performance of primarily female subjects with an age range of 44 to 56 years, ~24 to 36 years older than the subjects tested herein. Unlike the present study, he administered a cumulative dose of fluid lecithin (100 ml 9 g) or placebo in an alcohol/ water suspension (16.4 ml alcohol) in the morning (30 ml) and evening (60 ml) of the day between testing sessions over a 9-week period. Sorgatz did not report the exact dose of the PCh in the lecithin. Nor did he report the amount of lecithin or placebo consumed each day of the experiment for the taste analysis. The effects of alcohol on cerebral performance were not considered because the alcohol content was identical in the lecithin and placebo preparations. One interpretation of Sorgatz's findings is that a cumulative dose of PCh uniformly increases memory performance. Another possible interpretation is that the two variables of alcohol and subjects' age may have combined to produce a very slight decline in endogenous choline (26,27). Impairment in the cholinergic system either through pharmacological or surgical manipulation prior to the application of nootropic drugs has been used as the design model in animal research (28). This model has recently been employed with alcohol. Hodges et al. (29) has demonstrated that alcohol treated rats show improvement in learning after treatment with cholinergic agonists and disruption with antagonists at low doses that did not affect controls. Alcohol-induced cognitive impairments might be used as an effective independent variable in future PCh and human memory research. Unfortunately, the study of Harris et al. (10) also did not specify the purity of the lecithin administered, and, therefore, the dose of PCh is ambiguous. Given the current findings, a 10-g dose would not have produced a measurable treatment effect. Davis et al. (7), using 16 g of choline chloride in a cumulative dose, found insignificance in tests of digit span, memory retrieval, and memory storage. Neither Harris (10) nor Davis (7) analyzed the data based on learning rate. Therefore, either the subject variable of learning rate or the dose level may have contributed to the insignificant findings. Although Mohs et al. (8) also found insignificant results using 8 g of choline chloride in a cumulative dose, they reported a trend for lower baseline scores to be associated with better memory performance.

The practice of reporting only the dose of the substance administered, without conversion into its pure PCh or choline equivalent, makes it difficult to compare ACh precursors. For example, the choline in 10 g of choline chloride is not equivalent to the choline in 10 g of PCh, and the metabolic rates of choline chloride and PCh differ (19), which is why the cumulative doses of choline chloride given by Davis et al. (7) and Mohs et al. (8) are not comparable to the cumulative dose of PCh given by Sorgatz (9). Thirty minutes after the ingestion of choline chloride, serum-choline concentrations begin to decrease and continue to decrease over a 12 h period; serum-choline concentrations following lecithin ingestion begin to rise at 30 min, continue to increase until 8 h, and remain elevated for the 12 h measured (19). Because different researchers use different drugs to increase ACh, reporting the amount of choline supplied by the various pharmacological agents would facilitate comparisons between studies.

The results reported by Crook et al. (30) on the positive effects of phosphatidylserine, a precursor to PCh (31), on age-associated memory impairment and research demonstrating that large doses of glucose (32) can enhance memory performance presumably by augmentation of cholinergic function are consistent with our findings.

Further clinical studies, in which subject selection is based on reliable operational definitions of learning rate using a variety of memory assessment instruments, will be necessary to determine the precise dose-time curve and the specific subcomponents of memory that may be enhanced with PCh solutions. The usefulness of PCh, alone or in combination with other agents, on a given muscarinic receptor subtype in Alzheimer's disease should be investigated on the specific subcomponents of memory, particularly since recent research (33) with oral physostigmine, a cholinesterase inhibitor, has reported significant improvement in memory performance in Alzheimer's patients.

Given the current findings, popular claims that lecithin nutritional supplements are uniformly beneficial for memory problems may be misleading (34). Because the identification of a natural substance that will improve explicit memory on subgroups such as slow learners, the aged, or patients with neurological disorders involving memory loss has obvious benefits, more normative and clinical research specifying the dose-time parameters of PCh is warranted.

Acknowledgment: We thank Steve Blechman, Vice President, Research and Development, Twin Laboratories, Inc., for providing us with the PC-55 used in this research. We also are indebted to Dr. James Stevenson, Aerospace Human Factors Research Division, and to Dr. Joe Kamiya, Life Sciences Division, NASA-Ames Research Center, for their insightful comments. We thank Denny Burzynski and James Wilczak, Department of Mathematics, West Valley College, for their stimulating discussions. We are also grateful to Dr. John Gabrieli, Department of Psychology, Stanford University, for his constructive questions and advice.

Address correspondence and reprint requests to Prof. Sandra L. Ladd, Department of Psychology, West Valley College, 14000 Fruitvale Avenue, Saratoga, CA 95070-5698, U.S.A.

The preliminary study cited in this paper was presented at the 98th annual convention of the American Psychological Association, Clinical Neuropsychology Division, Boston, Massachusetts, August 10-14, 1990.

1. Beninger RJ, Wirsching BA, Jhamandas K, Boegamn RJ. Animal studies of brain acetylcholine and memory. Arch Gerontol Geriatr 1989;(suppl 1):71-89.

2. Olton DS, Wenk GL. Dementia: animal models of the cognitive impairments produced by degeneration of basal forebrain cholinergic system. In: Meltzer HY, ed. Psychopharmacology: the third generation of progress. New York: Raven Press, 1987:941-53.

3. Conway PG, Hamer RR. Future directions for drug development for the treatment of Alzheimer's disease. Clin Neuropharmacol 1991;14:v-vii.

4. Volger B. Alternatives in the treatment of memory loss in patients with Alzheimer's disease. Clin Pharm 1991;10:447-56.

5. Quirion R, Aubert 1, Lapchak PA, et al. Muscarinic receptor subtypes in human neurodegenerative disorders: focus on Alzheimer's disease. Trends Pharmacol Sci 1989;(Dec suppl):80-4.

6. Sitaram N, Weingartner H, Gillin JC. Human serial learning: enhancement with arecholine and choline and impairment with scopolamine. Science 1978;201:274-6.

7. Davis KL, Mohs RC, Tinklenberg JR, Hollister LE, Pfefferbaum A, Kopell BS. Cholinomimetics and memory: effect of choline chloride. Arch Neurol 1980;37:49-51.

8. Mohs RC, Davis KL, Tinklenberg JR, Hollister LE. Choline chloride effects on memory in the elderly. Neurobiol Aging 1980; 1: 21-5.

9. Sorgatz H. Effects of lecithin on memory and learning. In: Hanin 1, Ansell GB, eds. Lecithin: technological, biological, and therapeutic aspects. Proceedings of the fourth international colloquium on lecithin. New York: Plenum Press, 1987:147-53.

10. Harris CM, Dysken MW, Fovall P, Davis JM. Effect of lecithin on memory in normal adults. Am J Psychiatry 1983; 140:1010-12.

11. Gillin JC, Sitaram N, Weingartner H, Crammer C, Hanin 1. Effects of lecithin on memory and plasma choline levels: a study in normal volunteers. In: Pepeu G, Ladinsky H, eds. Cholinergic mechanisms: phylogenetic aspects, central and peripheral synapses and clinical significance. New York: Plenum Press, 1981:937-45.

12. Davis KL, Mohs RC, Tinklenberg JR, Pfefferbaum A. Physostigmine: improvement of long-term memory processes in normal humans. Science 1978;201:272-4.

13. Smith CM, Coogan JS, Hart S. The effects of physostigmine on memory test performance in normal volunteers. Psychopharmacology 1986;90:364-6.

14. Kopelman MD, Corn TH. Cholinergic 'blockade' as a model for cholinergic depletion. Brain 1988; 111: 1079-1110.

15. Rusted JM, Warburton DM. The effect of scopolamine on working memory in healthy young volunteers. Psychopharmacology 1988;96:145-52.

16. Cohen EL, Wurtman RJ. Brain acetylcholine: increase after systemic choline administration. Life Sci 1975;16:1095-1102.

17. Haubrich DR, Wang PFL, Clody DE, Wedeking PW. Increase in rat brain acetylcholine induced by choline or deanol. Life Sci 1975;17:975-80.

18. Wurtman RJ, Fernstrom JD. Control of brain neurotransmitter synthesis by precursor availability and nutritional state. Biochem Pharmacol 1976;25:1691-6.

19. Wurtman RJ, Hirsch MJ, Growdon JH. Lecithin consumption raises serum-free-choline levels. Lancet 1977;2:68-9.

20. Noble CE. Measurements of association value (a), rated associations (a'), and scaled meaningfulness (in') for the 2100 CVC combinations of the English alphabet. PsycholRep 1961;8:487-521.