- Email: [email protected]
Memory formation: A new three-phase model
Neuroscience Letters, 2 (1976) 165--169
165
© Elsevier/North-Holland, Amsterdam -- Printed in The Netherlands
MEMORY FORMATION: A NEW THREE-PHASE MODEL
MARIE E. GIBBS and KIM T. NG
Department of Psychology, La Trobe University, Bundoora, Vic. 3083 (Australia) (Received March 26th, 1976) (Accepted March 26th, 1976)
SUMMARY
A new three-phase pharmacological model of memory formation is proposed: (i) a short-term memory phase lasting 10 min, inhibited by lithium chloride (LiC1) and potassium chloride (KCI), but not by ouabain; (ii) a labile, sodium pump-dependent phase occurring after 10 min, inhibited by ouabain but not by cycloheximide (CXM); and (iii) a permanent, protein synthesis-dependent phase occurring after 30 min and inhibited by CXM. Diphenylhydantoin (DPH), which stimulates Na*-K÷-ATPase activity, overcomes ouabain- and CXM-induced amnesia but not LiC1- or KCl-induced amnesia.
Evidence from pharmacological studies with chicks has been interpreted to date [8,10] as supporting a two-phase model of memory: (i) an initial 'labile' phase lasting up to 30 min, and thought to be based on sodium pump activity because it can be inhibited by ouabain, copper and lithium; and (ii) a second, permanent phase, dependent on protein synthesis, which is inhibited by cycloheximide (CXM) [5]. We now present data in support of a new, three-phase model of memory. As in the standard procedure described elsewhere [2], we trained 1-day-old white-Leghorn black-Australorp cockerels to peck at a 4 mm chromed bead dipped in water, presented for 10 sec. Each chick was then given simple avoidance training by presenting the bead, after it was dipped in methyl anthranilate (NH~ C6 H4 COOCH3 ), for 10 sec in a single negative learning trial. Five minutes before the learning trial, a solution of saline (NaCI, 0.9%), lithium chloride (LiC1, 0.154 M), potassium chloride (KC1, 1.0 mM) or ouabain (0.4 #g/chick) was injected intracranially. The bilateral site of each 10 ~1 freehand injection was the centre of the forebrain at a depth of 3.0 mm. In the case of ouabain, a different group of chicks was also treated 15 min before learning. The chicks were given 0.1 ml of diphenylhydantoin (DPH, 0.1 mM) subcutaneously 5 rain after the learning trial. All drugs were freshly prepared in sterile NaC1 (0.9%, 0.154 M).
166 Retention tests using an identical chromed bead not dipped in the aversive solution were carried out at 5, 10, 15, 20, 25, 30 and 180 min following the learning trial, with a different group of 20 chickens at each learning-retention interval. Retention was measured as the percentage of animals avoiding the test bead in a single 10 sec presentation. The results from birds pretreated with NaC1, LiC1, KC1 or ouabain followed by post-treatment with DPH were compared with those from control birds receiving only pretreatment with NaC1, LiC1, KC1, or ouabain respectively. With LiC1 and KC1 pretreatments, amnesia is present 5 min after learning and the percentage of chicks avoiding the test bead is significantly lower than that found with saline pretreatment (X2 = 7.62 and 14.14 respectively in a X2 contingency test with degrees of freedom = 1, P<0.05 in each case). That the chicks learned the avoidance task is seen in that pecking of the methyl anthranilate bead resulted in reactions characteristic of aversion (e.g., violent shaking of the head and wiping of the beak) during the 10 sec learning trial and they did not repeck the bead thereafter in the learning period. Ouabain-induced amnesia does not occur until 15 min after learning, independent of whether the drug was administered 5 or 15 min before learning (X2 = 0.97 and 1.12 respectively for ouabain injected at --15 min and tested at 5 and 10 min; P > 0.05 in each case. X2 = 1.12 and 0.12 respectively for ouabain injected at --5 min and tested at 5 and 10 min; P > 0.05 in each case). CXM pretreatment induces amnesia after 30 min following learning, as reported previously [8,10]. Retention scores at 24 h with all four drugs were similar to those obtained at 180 min. The difference between the two early phases of m e m o r y storage is shown up clearly by the action of DPH, the results of which are shown in Fig. 2. DPH is effective in overcoming ouabain-induced amnesia, the percentage of chicks showing avoidance being comparable to that obtained with saline pretreated animals. DPH also successfully counteracts CXM-induced amnesia [6] This is consistent with the hypothesis [8,10] that ouabain induces amnesia through inhibition of sodium pump activity as DPH has been shown to stimulate Na÷-K÷-ATPase activity [6]. But DPH is ineffective in overcoming LiC1- and KCl-induced amnesia, the percentage of chicks avoiding being comparable to those pretreated with LiC1 and KC1, respectively, at any learning-retention interval. Subcutaneous injection of saline 5 min after learning did not produce the same effect as DPH. Therefore, any physiological reaction resulting from the injection procedure itself neither increased memory in the saline-pretreated chicks nor overcame amnesia in CXM pretreated chicks [6]. We interpret these findings as supporting the idea that a phase of m e m o r y lasting approximately 10 min must precede the labile phase postulated earlier [8,10]. The possibility that the previous finding of an absence of ouabain-induced amnesia at 10 min after learning is due to the time required by ouabain to act is ruled out by the present observation that ouabain administered either 15 or 5 min before learning did not induce amnesia at 10 min
167
100-
OUABAIN
B0-
60-
Z.0
20-
O Z Q
,
i
5 100"
,
//
KCI
LiCI
8060" /.0" 20"
lb
2'0
~o //-'180
, //---~
TIME OF TESTING (min)
Fig. 1. Percentage of chicks avoiding the test lure after pretreatment with saline, ouabain, LiC1 or KCI. Five minutes before learning chicks were given intracranial injections of saline (NaCl, 0.9%), LiCl ( 0 . 1 5 4 M), KCI (1.0 raM), or ouabain (0.4 rag/chick), In the case of ouabain, another group of chickens ( - ) were administered the drug 15 rain before learning. Different groups of chicks were used at each learning-retention interval.
after learning. Both times of administration lead to amnesia 15 min after learning. It may be noted that it was previously considered that lithium was inhibiting sodium pump activity [ 5 ] , but the present results suggest that it has at least an additional effect prior to the labile phase becoming active. Of interest is the report that in mammalian sympathetic ganglia, ganglionic transmission, although initially supported, is abolished during perfusion with lithium ions (0.154 M) [ 9 ] . Lithium may inhibit ACh release and/or synthesis in the cat cerebral cortex as well as interfering with conduction of impulses [ 1 ] Dose-response studies with lithium and ouabain [10] are consistent with the findings reported here. With increasing doses of LiC1 the percentage of chicks avoiding at the 10 min retention test decreases. With increasing doses of ouabain there is no change in the 10 min retention score. Further supportive evidence comes from the finding [4] that good discrimination between two coloured beads was observed when only one of the beads was made avervise on the learning trial. The retention present 10 min after learning in the
168
1°°7
OUABAIN*DPH
, //
z
'f'
'/"~
100-
'1'
LiCI + DPH
KCI + DPH
80"
60-
40-
20-
4' 10 DPH
20
,/J 30
,/,I 180
~ 1'0
DPH
2'0
30
1~)
TIME OF TESTING (rain)
Fig. 2. Percentage of chicks avoiding the test lure after post-treatment with DPH. Saline, ouabain, LiCI and KC1 pretreated chicks were given subcutaneous injections of DPH (0.1 raM) 5 rain after learning. Different groups of chicks were used at each learning retention interval.
presence of ouabain is not a generalized avoidance response but specific avoidance of the particular characteristic.s of the bead used in learning. In successive presentation of t w o beads 10 min after learning, chicks continue to peck the neutral bead but avoid the bead which is aversive on the learning trial. The fact that DPH fails to counteract LiC1- and KCl-induced amnesia has significant implications for the role of the second, labile phase of memory. It supports the previous hypothesis that sodium pump activity serves the function of maintaining the 10 min short-term memory trace for a sufficient period of time to allow protein synthesis to take place. This phase is most likely to be involved in the biochemical consequences of reinforcement [2,3,7] If the short-term (10 rain) trace is inhibited, as in the case with the administration of LiC1 or KC1, the activity of the sodium pump cannot be instigated. In this case enhancement of Na÷-K÷-ATPase activity by DPH has no further
169
consequences for maintaining memory. While the present nature of the initial short-term trace is still unknown, the fact that it is inhibited by LiCI and KC1 suggests that the memory might be held by some mechanism involving nervous impulse activity. T h e subsequent labile phase, susceptible to ouabain and restored by DPH is now cast in the role of maintaining the consequences of impulse activity (activation of Na+-K+-ATPase) until they ar~ coded into more permanent traces through protein synthesis. This threephase model based on speculations about the basic physical nature of memory storage is surprisingly consistent with models of memory storage which have come out of cognitive studies in humans, at least with regard to time factors. REFERENCES 1 Bjegovic, M. and Randic, M., Effect of lithium ions on the release of acetylcholine from the cerebral cortex, Nature (Lond.), 230 (1971) 587--588. 2 Gibbs, M.E., The effects of amphetamine on short-term, protein-independent, memory in day-old chickens, Pharmacol. Biochem. Behav., (1976) in press. 3 Gibbs, M.E., Modulation of cycloheximide-resistant memory by sympathomimetic agents, Pharmacol. Biochem. Behav., (1976) in press. 4 Gibbs, M.E., and Barnett, J.M., Drug effects on successive discrimination learning in young chickens, Brain Res. Bull., (1976) in press. 5 Gibbs, M.E., Jeffrey, P.L., Austin, L. and Mark, R.F., Separate biochemical actions of inhibitors of short- and long-term memory, Pharmacol. Biochem. Behav., 1 (1973) 693--701. 6 Gibbs, M.E. and Ng, K.T., Diphenylhydantoin facilitation of labile, protein-independent memory, Brain Res. Bull.,(1976) in press. 7 Jeffrey, P.L. and Gibbs, M.E., Biochemical actions of sympathomimetic drugs on protein-independent, labile memory, submitted. 8 Mark, R.F. and Watts, M.E., Drug inhibition of m e m o r y formation in chickens. I. Long-term memory, Proc. roy. Soc. B, 178 (1971) 439--454. 9 Pappano, A.J. and Voile, R.L., Actions of lithium ions in mammalian sympathetic ganglia, J. Pharmacol. exp. Ther., 157 (1967) 346--355. 10 Watts, M.E. and Mark, R.F., Drug inhibition of m e m o r y formation in chickens. II. Short-term memory, Proc. roy. Soc. B, 178 (1971) 455--464.
165
© Elsevier/North-Holland, Amsterdam -- Printed in The Netherlands
MEMORY FORMATION: A NEW THREE-PHASE MODEL
MARIE E. GIBBS and KIM T. NG
Department of Psychology, La Trobe University, Bundoora, Vic. 3083 (Australia) (Received March 26th, 1976) (Accepted March 26th, 1976)
SUMMARY
A new three-phase pharmacological model of memory formation is proposed: (i) a short-term memory phase lasting 10 min, inhibited by lithium chloride (LiC1) and potassium chloride (KCI), but not by ouabain; (ii) a labile, sodium pump-dependent phase occurring after 10 min, inhibited by ouabain but not by cycloheximide (CXM); and (iii) a permanent, protein synthesis-dependent phase occurring after 30 min and inhibited by CXM. Diphenylhydantoin (DPH), which stimulates Na*-K÷-ATPase activity, overcomes ouabain- and CXM-induced amnesia but not LiC1- or KCl-induced amnesia.
Evidence from pharmacological studies with chicks has been interpreted to date [8,10] as supporting a two-phase model of memory: (i) an initial 'labile' phase lasting up to 30 min, and thought to be based on sodium pump activity because it can be inhibited by ouabain, copper and lithium; and (ii) a second, permanent phase, dependent on protein synthesis, which is inhibited by cycloheximide (CXM) [5]. We now present data in support of a new, three-phase model of memory. As in the standard procedure described elsewhere [2], we trained 1-day-old white-Leghorn black-Australorp cockerels to peck at a 4 mm chromed bead dipped in water, presented for 10 sec. Each chick was then given simple avoidance training by presenting the bead, after it was dipped in methyl anthranilate (NH~ C6 H4 COOCH3 ), for 10 sec in a single negative learning trial. Five minutes before the learning trial, a solution of saline (NaCI, 0.9%), lithium chloride (LiC1, 0.154 M), potassium chloride (KC1, 1.0 mM) or ouabain (0.4 #g/chick) was injected intracranially. The bilateral site of each 10 ~1 freehand injection was the centre of the forebrain at a depth of 3.0 mm. In the case of ouabain, a different group of chicks was also treated 15 min before learning. The chicks were given 0.1 ml of diphenylhydantoin (DPH, 0.1 mM) subcutaneously 5 rain after the learning trial. All drugs were freshly prepared in sterile NaC1 (0.9%, 0.154 M).
166 Retention tests using an identical chromed bead not dipped in the aversive solution were carried out at 5, 10, 15, 20, 25, 30 and 180 min following the learning trial, with a different group of 20 chickens at each learning-retention interval. Retention was measured as the percentage of animals avoiding the test bead in a single 10 sec presentation. The results from birds pretreated with NaC1, LiC1, KC1 or ouabain followed by post-treatment with DPH were compared with those from control birds receiving only pretreatment with NaC1, LiC1, KC1, or ouabain respectively. With LiC1 and KC1 pretreatments, amnesia is present 5 min after learning and the percentage of chicks avoiding the test bead is significantly lower than that found with saline pretreatment (X2 = 7.62 and 14.14 respectively in a X2 contingency test with degrees of freedom = 1, P<0.05 in each case). That the chicks learned the avoidance task is seen in that pecking of the methyl anthranilate bead resulted in reactions characteristic of aversion (e.g., violent shaking of the head and wiping of the beak) during the 10 sec learning trial and they did not repeck the bead thereafter in the learning period. Ouabain-induced amnesia does not occur until 15 min after learning, independent of whether the drug was administered 5 or 15 min before learning (X2 = 0.97 and 1.12 respectively for ouabain injected at --15 min and tested at 5 and 10 min; P > 0.05 in each case. X2 = 1.12 and 0.12 respectively for ouabain injected at --5 min and tested at 5 and 10 min; P > 0.05 in each case). CXM pretreatment induces amnesia after 30 min following learning, as reported previously [8,10]. Retention scores at 24 h with all four drugs were similar to those obtained at 180 min. The difference between the two early phases of m e m o r y storage is shown up clearly by the action of DPH, the results of which are shown in Fig. 2. DPH is effective in overcoming ouabain-induced amnesia, the percentage of chicks showing avoidance being comparable to that obtained with saline pretreated animals. DPH also successfully counteracts CXM-induced amnesia [6] This is consistent with the hypothesis [8,10] that ouabain induces amnesia through inhibition of sodium pump activity as DPH has been shown to stimulate Na÷-K÷-ATPase activity [6]. But DPH is ineffective in overcoming LiC1- and KCl-induced amnesia, the percentage of chicks avoiding being comparable to those pretreated with LiC1 and KC1, respectively, at any learning-retention interval. Subcutaneous injection of saline 5 min after learning did not produce the same effect as DPH. Therefore, any physiological reaction resulting from the injection procedure itself neither increased memory in the saline-pretreated chicks nor overcame amnesia in CXM pretreated chicks [6]. We interpret these findings as supporting the idea that a phase of m e m o r y lasting approximately 10 min must precede the labile phase postulated earlier [8,10]. The possibility that the previous finding of an absence of ouabain-induced amnesia at 10 min after learning is due to the time required by ouabain to act is ruled out by the present observation that ouabain administered either 15 or 5 min before learning did not induce amnesia at 10 min
167
100-
OUABAIN
B0-
60-
Z.0
20-
O Z Q
,
i
5 100"
,
//
KCI
LiCI
8060" /.0" 20"
lb
2'0
~o //-'180
, //---~
TIME OF TESTING (min)
Fig. 1. Percentage of chicks avoiding the test lure after pretreatment with saline, ouabain, LiC1 or KCI. Five minutes before learning chicks were given intracranial injections of saline (NaCl, 0.9%), LiCl ( 0 . 1 5 4 M), KCI (1.0 raM), or ouabain (0.4 rag/chick), In the case of ouabain, another group of chickens ( - ) were administered the drug 15 rain before learning. Different groups of chicks were used at each learning-retention interval.
after learning. Both times of administration lead to amnesia 15 min after learning. It may be noted that it was previously considered that lithium was inhibiting sodium pump activity [ 5 ] , but the present results suggest that it has at least an additional effect prior to the labile phase becoming active. Of interest is the report that in mammalian sympathetic ganglia, ganglionic transmission, although initially supported, is abolished during perfusion with lithium ions (0.154 M) [ 9 ] . Lithium may inhibit ACh release and/or synthesis in the cat cerebral cortex as well as interfering with conduction of impulses [ 1 ] Dose-response studies with lithium and ouabain [10] are consistent with the findings reported here. With increasing doses of LiC1 the percentage of chicks avoiding at the 10 min retention test decreases. With increasing doses of ouabain there is no change in the 10 min retention score. Further supportive evidence comes from the finding [4] that good discrimination between two coloured beads was observed when only one of the beads was made avervise on the learning trial. The retention present 10 min after learning in the
168
1°°7
OUABAIN*DPH
, //
z
'f'
'/"~
100-
'1'
LiCI + DPH
KCI + DPH
80"
60-
40-
20-
4' 10 DPH
20
,/J 30
,/,I 180
~ 1'0
DPH
2'0
30
1~)
TIME OF TESTING (rain)
Fig. 2. Percentage of chicks avoiding the test lure after post-treatment with DPH. Saline, ouabain, LiCI and KC1 pretreated chicks were given subcutaneous injections of DPH (0.1 raM) 5 rain after learning. Different groups of chicks were used at each learning retention interval.
presence of ouabain is not a generalized avoidance response but specific avoidance of the particular characteristic.s of the bead used in learning. In successive presentation of t w o beads 10 min after learning, chicks continue to peck the neutral bead but avoid the bead which is aversive on the learning trial. The fact that DPH fails to counteract LiC1- and KCl-induced amnesia has significant implications for the role of the second, labile phase of memory. It supports the previous hypothesis that sodium pump activity serves the function of maintaining the 10 min short-term memory trace for a sufficient period of time to allow protein synthesis to take place. This phase is most likely to be involved in the biochemical consequences of reinforcement [2,3,7] If the short-term (10 rain) trace is inhibited, as in the case with the administration of LiC1 or KC1, the activity of the sodium pump cannot be instigated. In this case enhancement of Na÷-K÷-ATPase activity by DPH has no further
169
consequences for maintaining memory. While the present nature of the initial short-term trace is still unknown, the fact that it is inhibited by LiCI and KC1 suggests that the memory might be held by some mechanism involving nervous impulse activity. T h e subsequent labile phase, susceptible to ouabain and restored by DPH is now cast in the role of maintaining the consequences of impulse activity (activation of Na+-K+-ATPase) until they ar~ coded into more permanent traces through protein synthesis. This threephase model based on speculations about the basic physical nature of memory storage is surprisingly consistent with models of memory storage which have come out of cognitive studies in humans, at least with regard to time factors. REFERENCES 1 Bjegovic, M. and Randic, M., Effect of lithium ions on the release of acetylcholine from the cerebral cortex, Nature (Lond.), 230 (1971) 587--588. 2 Gibbs, M.E., The effects of amphetamine on short-term, protein-independent, memory in day-old chickens, Pharmacol. Biochem. Behav., (1976) in press. 3 Gibbs, M.E., Modulation of cycloheximide-resistant memory by sympathomimetic agents, Pharmacol. Biochem. Behav., (1976) in press. 4 Gibbs, M.E., and Barnett, J.M., Drug effects on successive discrimination learning in young chickens, Brain Res. Bull., (1976) in press. 5 Gibbs, M.E., Jeffrey, P.L., Austin, L. and Mark, R.F., Separate biochemical actions of inhibitors of short- and long-term memory, Pharmacol. Biochem. Behav., 1 (1973) 693--701. 6 Gibbs, M.E. and Ng, K.T., Diphenylhydantoin facilitation of labile, protein-independent memory, Brain Res. Bull.,(1976) in press. 7 Jeffrey, P.L. and Gibbs, M.E., Biochemical actions of sympathomimetic drugs on protein-independent, labile memory, submitted. 8 Mark, R.F. and Watts, M.E., Drug inhibition of m e m o r y formation in chickens. I. Long-term memory, Proc. roy. Soc. B, 178 (1971) 439--454. 9 Pappano, A.J. and Voile, R.L., Actions of lithium ions in mammalian sympathetic ganglia, J. Pharmacol. exp. Ther., 157 (1967) 346--355. 10 Watts, M.E. and Mark, R.F., Drug inhibition of m e m o r y formation in chickens. II. Short-term memory, Proc. roy. Soc. B, 178 (1971) 455--464.