Abstract
5-Hydroxytryptamine (5-HT) transmission has been implicated in memory and in depression. Both 5-HT depletion and specific 5-HT agonists lower memory performance, while depression is also associated with memory deficits. The precise neuropharmacology and neural mechanisms underlying these effects are unknown. We used neural network simulations to elucidate the neuropharmacology and network mechanisms underlying 5-HT effects on memory. The model predicts that these effects are largely dependent on transmission over the 5-HT1A and 5-HT3 receptors, which regulate the selectivity of retrieval. It also predicts differential memory deficit profiles for 5-HT depletion and overactivation. The latter predictions were confirmed in studies with healthy and depressed participants undergoing acute tryptophan depletion or ipsipirone challenge. The results suggest that the memory impairments in depressed subjects may be related to 5-HT undertransmission, and support the notion that 5-HT1A agonists ameliorate memory deficits in depression.
Similar content being viewed by others
Log in or create a free account to read this content
Gain free access to this article, as well as selected content from this journal and more on nature.com
or
References
Abercrombie HC, Kalin NH, Thurow ME, Rosenkranz MA, Davidson RJ (2003). Cortisol variation in humans affects memory for emotionally laden and neutral information. Behav Neurosci 117: 505–516.
Acsády L, Halasy K, Freund TF (1993). Calretinin is present in non-pyramidal cells of the rat hippocampus-III. Theri inputs from the median raphe and medial septal nuclei. Neuroscience 52: 829–841.
Allain H, Lieuniy A, Brunet BF, Mirabaud C, Trebon P, LeCoz F et al (1992). Antidepressants and cognition: comparative effects of moclobemide, viloxazine and maprotiline. Psychopharmacol Suppl 106: 58–61.
Altman HJ, Normile HJ (1988). What is the nature of the role of the serotonergic nervous system in learning and memory: prospects for development of an effective treatment strategy for senile dementia. Neurobiol Aging 9: 627–638.
Anderson GM, Barr CS, Lindell S, Durham AC, Shifrovich I, Higley JD (2005). Time course of the effects of the serotonin-selective reuptake inhibitor sertraline on central and peripheral serotonin neurochemistry in the rhesus monkey. Psychopharmacology (Berlin) 178: 339–346.
Andrade R, Chaput Y (1991). 5-Hydroxytrytamine4-like receptors mediate the slow excitatory response to serotonin in the rat hippocampus. J Pharmacol Exp Ther 257: 930–937.
Andrade R, Nicoll RA (1987). Pharmacologically distinct actions of serotonin in the rat hippocampus. J Physiol 394: 99–124.
Bacon WL, Beck SG (2000). 5-Hydroxytryptamine(7) receptor activation decreases slow afterhyperpolarization amplitude in CA3 hippocampal pyramidal cells. J Pharmacol Exp Ther 294: 672–679.
Banks WP (1970). Signal detection theory and human memory. Psychol Bull 74: 81–99.
Barnes NM, Sharp T (1999). A review of central 5-HT receptors and their function. Neuropharmacology 38: 1083–1152.
Buhot MC, Martin S, Segu L (2000). Role of serotonin in memory impairment. Ann Med 32: 210–221.
Burnet PWJ, Eastwood SL, Lacey K, Harrison PJ (1995). The distribution of 5-HT1A and 5-HT2A receptor mRNA in human brain. Brain Res 676: 157–168.
Dong J, de Montigny C, Blier P (1997). Effect of acute and repeated versus sustained administration of the 5-HT1A receptor agonist ipsapirone: electrophysiological studies in the rat hippocampus and dorsal raphe. Naunyn-Schmiedeberg Arch Pharmacol 356: 303–311.
Dragoi G, Carpi M, Recce M, Csicsvari J, Buzsaki G (1999). Interactions between hippocampus and medial septum during sharp wave and theta oscillation in the behaving rat. J Neurosci 19: 6191–6199.
Fleming SK, Blasey C, Schatzberg AF (2004). Neuropsychological correlates of psychotic features in major depressive disorders: a review and meta-analysis. J Psychiatr Res 38: 27–35.
Freund T, Antal M (1988). GABA-containing neurons in the septum control inhibitory interneurons in the hippocampus. Nature 336: 170–173.
Fudge JL, Perry PJ, Garvey MJ, Kelly MW (1990). A comparision of the effect of fluoxetine and trazodone on the cognitive functioning of depressed outpatients. J Affective Disorders 18: 275–280.
Glanzer M, Adams JK (1985). The mirror effect in recognition memory. Mem Cognit 13: 8–20.
Glanzer M, Adams JK, Iverson GJ, Kim K (1993). The regularities of recognition memory. Psychol Rev 100: 546–567.
Gluck MA, Meeter M, Myers CE (2003). Computational models of the hippocampal region: linking incremental learning and episodic memory. Trends Cogn Sci 7: 269–276.
Gulyás AI, Görcs TJ, Freund TF (1990). Innervation of different peptide-containing neurons in the hippocampus by GABAergic septal afferents. Neuroscience 37: 31–44.
Gulyás AI, Seress L, Toth K, Acsady L, Antal M, Freund TF (1991). Septal GABAergic neurons innervate inhibitory interneurons in the hippocampus of the macaque monkey. Neuroscience 41: 381–390.
Harmer CJ, Bhagwagar Z, Cowen PJ, Goodwin GM (2002). Acute administration of citalopram facilitates memory consolidation in healthy volunteers. Psychopharmacology (Berlin) 163: 106–110.
Hasselmo ME (1995). Neuromodulation and cortical function: modeling the physiological basis of behavior. Behav Brain Res 67: 1–27.
Hasselmo ME, Bower JM (1993). Acetylcholine and memory. Trends Neurosci 16: 218–222.
Hasselmo ME, Fehlau BP (2001). Differences in time course of ACh and GABA modulation of excitatory synaptic potentials in slices of rat hippocampus. J Neurophysiol 86: 1792–1802.
Hasselmo ME, Schnell E, Barkai E (1995). Dynamics of learning and recall at excitatory recurrent synapses and cholinergic modulation in rat hippocampal region CA3. J Neurosci 15: 5249–5262.
Hodgkin AL, Huxley AF (1952). A quantitative description of ion currents and its application to conduction and excitation in nerve membranes. J Physiol 117: 500–544.
Humphreys MS, Bain JD, Pike R (1989). Different ways to cue a coherent system: a theory for episodic, semantic and procedural tasks. Psychol Rev 96: 208–233.
Jagannathan V, Venitz J (1997). Pharmacokinetics and CNS pharmacodynamics of the 5-HT1A agonist buspirone in humans following acute L-tryptophan depletion challenge. Methods Find Exp Clin Pharmacol 19: 351–361.
Jelicic M, Geraerts E, Merckelbach H, Guerrieri R (2004). Acute stress enhances memory for emotional words, but impairs memory for neutral words. Int J Neurosci 114: 1343–1351.
Johnson MH, Magaro PA (1987). Effects of mood and severity on memory processes in depression and mania. Psychol Bull 101: 28–40.
Levkovitz Y, Caftori R, Avital A, Richter-Levin G (2002). The SSRIs drug Fluoxetine, but not the noradrenergic tricyclic drug Desipramine, improves memory performance during acute major depression. Brain Res Bull 15: 345–350.
MacGregor RJ, Oliver RM (1974). A model for repetitive firing in neurons. Cybernetik 16: 53–64.
McLennan H, Miller JJ (1974). The hippocampal control of neuronal discharges in the septum of the rat. J Physiol 237: 607–624.
Meeter M, Talamini LM, Murre JMJ (2004). Mode shifting between storage and recall based on novelty detection in oscillating hippocampal circuits. Hippocampus 14: 722–741.
Meltzer HY, Maes M (1995). Effects of ipsapirone on plasma cortisol and body temperature in major depression. Biol Psychiatry 38: 450–457.
Middlemiss DN, Price GW, Watson JM (2002). Serotonergic targets in depression. Curr Opin Pharmacol 2: 18–22.
Naughton M, Mulrooney JB, Leonard BE (2000). A review of the role of serotonin receptors in psyciatric disorders. Hum Psychopharmacol 15: 397–415.
Peroutka SJ (1988). 5-Hydroxytryptamine receptor subtypes: molecular, biochemical and physiological characterization. Trends Neurosci 11: 496–500.
Piercey MF, Smith MW, Lum-Ragan JT (1994). Excitation of noradrenergic cell firing by 5-hydroxytryptamine1A agonists correlates with dopamine antagonist properties. J Pharmacol Exp Ther 268: 1297–1303.
Piguet P, Galvan M (1994). Transient and long-lasting actions of 5-HT on rat dentate gyrus neurones in vitro. J Physiol 481: 629–639.
Riedel WJ (2004). Editorial: Cognitive changes after acute tryptophan depletion; what can they tell us? Psychol Med 34: 3–8.
Riedel WJ, Eikmans K, Heldens A, Schmitt JAJ (2005). Specific serotonergic reuptake inhibition impairs vigilance performance acutely and after subchronic treatment. J Psychopharmacol 19: 12–20.
Riedel WJ, Klaassen T, Deutz NEP, van Someren A, van Praag HM (1999). Tryptophan depletion in normal volunteers produces selective impairment in memory consolidation. Psychopharmacology 141: 362–369.
Riedel WJ, Klaassen T, Griez E, Honig A, Menheere PPCA, van Praag HM (2002). Dissociable hormonal, cognitive and mood responses to neuroendocrine challenge: evidence for receptor-specific serotonergic dysregulation in depressed mood. Neuropsychopharmacology 26: 358–367.
Ruiz JC, Soler MJ, Dasi C (2004). Study time effects in recognition memory. Percept Motor Skills 98: 638–642.
Schaub RT, Linden M, Copeland JRM (2003). A comparison of GMS-A/AGECAT, DSM-III-R for dementia and depression, including subthreshold depression (SD): results from the Berlin Aging Study (BASE). Int J Geriatr Psychiatry 18: 109–117.
Schmitt JAJ, Jorissen BL, Sobczak S, van Boxtel MPJ, Hogervorst E, Deutz NEP et al. (2000). Tryptophan depletion impairs memory consolidation but improves focused attention in healthy young volunteers. J Psychopharmacol 14: 21–29.
Schmitt JAJ, Kruizinga MJ, Riedel WJ (2001). Non-serotonergic pharmacological profiles and associated cognitive effects of serotonin reuptake inhibitors. J Psychopharmacol 15: 173–179.
Siegfried K, O'Connolly M (1986). Cognitive and psychomotor effects of different antidepressants in the treatment of old age depression. J Psychopharmacol 16: 207–214.
Snodgrass JG, Corwin J (1988). Pragmatics of measuring recognition memory: applications to dementia and amnesia. J Exp Psychol Gen 117: 34–50.
Sohal VS, Hasselmo ME (1998). GABA-b modulation improves sequence disambiguation in computational models of hippocampal region CA3. Hippocampus 8: 171–193.
Torres GE, Arfken CL, Andrade R (1996). 5-Hydroxytryptamine4 receptors reduce afterhyperpolarization in hippocampus by inhibiting calcium-induced calcium release. Mol Pharmacol 50: 1316–1322.
Tóth K, Borhegyi Z, Freund TF (1993). Postsynaptic targets of GABAergic hippocampal neurons in the medial septum-diagonal band of Broca complex. J Neurosci 13: 3712–3724.
Tóth K, Freund TF, Miles R (1997). Disinhibition of rat hippocampal pyramidal cells by GABAergic afferents from the septum. J Physiol 500: 463–474.
Witter MP, Wouterlood FG, Naber PA, Van Haeften T (2000). Anatomical organization of the parahippocampal–hippocampal network. Ann NY Acad Sci 911: 1–24.
Yatham LN, Steiner M (1993). Neuroendocrine probes of serotonergic function: a critical review. Life Sci 53: 447–463.
Acknowledgements
The research of M Meeter was supported by grant 402-01-630 from the Dutch National Science Foundation (NWO). We thank Arjan Blokland for stimulating discussions. The contribution of W Riedel and interpretation of research data for this article were entirely carried out in the University of Maastricht. Currently, W Riedel is also affiliated to GlaxoSmithKline R&D.
Author information
Authors and Affiliations
Corresponding author
Additional information
Supplementary Information accompanies the paper on the Neuropsychopharmacology website (http://www.nature.com/npp)
Supplementary information
Appendix
Appendix
Integrate-and-fire MacGregor model neurons were used for the model. In running the simulations, the discrete-time approximation formulas given by MacGregor and Oliver (1974) were used. The model was constructed using the Nutshell simulator, developed in our group. It can be downloaded without cost at www.neuromod.org/nutshell.
MacGregor and Oliver (1974) derived their model neuron from the Hodgkin–Huxley formulas (Hodgkin and Huxley, 1952) to account for firing characteristics in single neurons, while being computationally inexpensive enough for use in large-scale networks. These model neurons show spiking, adaptation, and threshold accommodation (the latter was not implemented in the present simulations). They are updated in discrete time steps, which in our simulations lasted 2 ms.
The model neuron emits a spike every time the membrane potential E crosses the threshold θ:
In this equation, S is a dichotomous variable that is equal to 1 if the node emits a spike, and equals 0 otherwise. The membrane potential, E, is dependent on the sodium, potassium, and chloride currents over the membrane, as described in the following differential equation:
Here, −δE is the leak current, gex the excitatory conductance, Eex the sodium reversal potential, gi the inhibitory conductance, and Ei the chloride reversal potential. For computational purposes, both the membrane potential and the reversal potentials were mapped onto the interval [−1, 7] via a simple linear transformation (MacGregor and Oliver, 1974). Resting potential is equated to 0 (−75 mV), the firing threshold θ to 1 (−60 mV), the sodium reversal potential to 7 (+30 mV), and both the potassium and chloride reversal potentials to −1 (−90 mV). The parameter governing the leak current, δ, is set to 1/7. When the node emits a spike, membrane potential is reset to resting level (via the term SE).
The potassium conductance gk models adaptation, and is determined by
where S is the spiking variable. The time constant τ is set to 1/13, the gain parameter b to 0.35. Excitatory input to the ith node is a simple linear summation of weighted inputs to that node:
where wij is the weight from node j to node i, and Sj is the spiking variable of node j. Rise times of synaptic inputs are thus not taken into account.
Simple Hebbian learning is used, modeling LTP, with the additions of negative Hebbian learning, modeling LTD, and a bound on connection weights. Weights are changed according to
Here, wij is the weight from node j to node i, while Si and Sj are the spiking variables of the receiving and sending node, respectively. This is subject to the constraints that a weight cannot be lower than 0 or exceed a maximum W. The positive learning rate, μ+, as well as the maximum weight, W, are set separately for every connection (see Table 1). The negative learning rate μ− is set to 75% of the positive learning rate in all connections.
The inhibitory conductance, gi, in a given layer, l, is modeled as a continuous variable reflecting firing rates of inhibitory interneurons. It is described by the following equation:
where st is the activity of the septal interneuron:
This is a simple sinusoid between 0 and 1 with a frequency of f (set to 50, equivalent to a 200 ms θ-band oscillation). The other component of Equation 6, itl, models the activity of intrinsic interneurons:
Thus, inhibition in layer l on time step t is a function of the feed-forward and feedback activation of inhibitory cells by the pyramidal cells, and of inhibition on time step t−1. Feed-forward and feedback inhibition are linear functions of the excitatory activation in the layers connecting to layer l (feed-forward), and of excitatory activation in layer l itself (feedback). The activation of each layer (Al) is calculated by dividing the number of firing nodes in the layer by its maximum kl (kEC=12, kDG=10, kCA3=10, kCA1=12). The βλ parameter (strength of feedback inhibition to layer l) was equal to 0.5 for EC, CA3, and CA1, and to 2 in layer DG. The λlp parameters associated with each connection (strength of feedforward inhibition from layer p to layer l) are listed in Table 1. No rise time is included in the formula for inhibition, as our 2 ms time step made this redundant. However, the decay parameter of the current (αι) was set by fitting a single exponential to the double exponential used by Sohal and Hasselmo (1998); αι=0.76.
In very large networks, the inhibition described above will be sufficient to constrain activity. In networks of the size used here, random fluctuations may produce large swings in activity that can be kept in check with a fast cutoff mechanism. This mechanism allows no more than a kl number of nodes to fire in a layer at any given time step. If more than kl nodes cross the firing threshold, only the kl nodes with the highest membrane potential are allowed to fire.
ACh levels in the model are regulated by inhibitory activity in layers CA3 and CA1. Activity of the septal cholinergic neurons, Ats, is set to F−inhibition (see Equation 9). Here, F, set to 1 in all simulations, represents excitation of the septum by sources external to the model, such as the reticular formation. Inhibition comes from the septal oscillator interneurons, st (whose output is the θ-frequency sinusoid given by Equation 7), and from the hippocampal afferents, its. A moving average of inhibition in CA1 and CA3 determines its (given by Equation 10).
The parameter αs is set to 0.85, and βs to 0.45. Release of ACh is equal to the activity of the septal cholinergic node, Ats. This release, in turn, determines ACh modulation in the hippocampus, for which we use the symbol Ψ, following Hasselmo et al (1995). At each time step, the amount of ACh released is fed into a dual exponential:
The time constants (τ1, τ2) of the dual exponential were rescaled from those found by Hasselmo and Fehlau (2001), who fitted a dual exponential to experimental data on the time course of ACh modulation data (τ1=0.001258, τ2=0.00015). These values correspond to a slow rise with a maximum at around 3.5 s, and a decrease back to 0 in 10–20 s.
As the effects of acetylcholine have been discussed in the main text, only their implementation will be listed here.
-
1)
For preferential dampening of transmission over Schaffer collaterals to CA3 and CA1, transmission in these two tracts (gex in Equation 4) is multiplied by a factor 1−0.6*Ψ.
-
2)
For enhancement of LTP at CA3 recurrent collateral synapses and at CA1 Schaffer collateral synapses, the learning rate (μ in Equation 5) is multiplied by Ψ in these connections.
-
3)
Reduction of firing adaptation of DG, CA3, and CA1 excitatory cells is effectuated by multiplication of the adaptation constant (b in Equation 3) with a factor 1−Ψ.
-
4)
Suppression of inhibition in all model layers is achieved multiplying the feedback inhibition constant (α in Equation 8) by a factor 1−0.5*Ψ.
-
5)
A mild depolarization of DG, CA3, and CA1 principle cells is implemented adding a constant factor, 0.12*Ψ, to the input of cells in these layers (gex in Equation 4).
Rights and permissions
About this article
Cite this article
Meeter, M., Talamini, L., Schmitt, J. et al. Effects of 5-HT on Memory and the Hippocampus: Model and Data. Neuropsychopharmacol 31, 712–720 (2006). https://doi.org/10.1038/sj.npp.1300869
Received:
Revised:
Accepted:
Published:
Issue date:
DOI: https://doi.org/10.1038/sj.npp.1300869
Keywords
This article is cited by
-
Neurobiological effects of aerobic exercise, with a focus on patients with schizophrenia
European Archives of Psychiatry and Clinical Neuroscience (2019)
-
The relationship between reward and punishment processing and the 5-HT1A receptor as shown by PET
Psychopharmacology (2014)
-
Molecular Docking of Bacosides with Tryptophan Hydroxylase: A Model to Understand the Bacosides Mechanism
Natural Products and Bioprospecting (2014)
-
Streptozotocin-induced insulin deficiency leads to development of behavioral deficits in rats
Acta Neurologica Belgica (2013)
-
Serotonin-mediated modulation of Na+/K+ pump current in rat hippocampal CA1 pyramidal neurons
BMC Neuroscience (2012)
