Theta oscillations in human cortex during a working-memory task: evidence for local generators.

Cortical theta appears important in sensory processing and memory. Intracanial electrode recordings provide a high spatial resolution method for studying such oscillations during cognitive tasks. Recent work revealed sites at which oscillations in the theta range (4-12 Hz) could be gated by a working-memory task: theta power was increased at task onset and continued until task offset. Using a large data set that has now been collected (10 participants/619 recording sites), we have sufficient sampling to determine how these gated sites are distributed in the cortex and how they are synchronized. A substantial fraction of sites in occipital/parietal (45/157) and temporal (23/280) cortices were gated by the task. Surprisingly, this aspect of working-memory function was virtually absent in frontal cortex (2/182). Coherence measures were used to analyze the synchronization of oscillations. We suspected that because of their coordinate regulation by the working-memory task, gated sites would have synchronized theta oscillations. We found that, whereas nearby gated sites (<20 mm) were often but not always coherent, distant gated sites were almost never coherent. Our results imply that there are local mechanisms for the generation of cortical theta.

Storage, recall, and novelty detection of sequences by the hippocampus: elaborating on the SOCRATIC model to account for normal and aberrant effects of dopamine.

In order to understand how the molecular or cellular defects that underlie a disease of the nervous system lead to the observable symptoms, it is necessary to develop a large-scale neural model. Such a model must specify how specific molecular processes contribute to neuronal function, how neurons contribute to network function, and how networks interact to produce behavior. This is a challenging undertaking, but some limited progress has been made in understanding the memory functions of the hippocampus with this degree of detail. There is increasing evidence that the hippocampus has a special role in the learning of sequences and the linkage of specific memories to context. In the first part of this paper, we review a model (the SOCRATIC model) that describes how the dentate and CA3 hippocampal regions could store and recall memory sequences in context. A major line of evidence for sequence recall is the "phase precession" of hippocampal place cells. In the second part of the paper, we review the evidence for theta-gamma phase coding. According to a framework that incorporates this form of coding, the phase precession is interpreted as cued recall of a discrete sequence of items from long-term memory. The third part of the paper deals with the issue of how the hippocampus could learn memory sequences. We show that if multiple items can be active within a theta cycle through the action of a short-term "buffer," NMDA-dependent plasticity can lead to the learning of sequences presented at realistic item separation intervals. The evidence for such a buffer function is reviewed. An important underlying issue is whether the hippocampal circuitry is configured differently for learning and recall. We argue that there are indeed separate states for learning and recall, but that both involve theta oscillations, albeit in possibly different forms. This raises the question of how neuromodulatory input might switch the hippocampus between learning and recall states and more generally how different neuromodulatory inputs reconfigure the hippocampus for different functions. In the fifth part of this paper we review our studies of dopamine and dopamine/NMDA interactions in the control of synaptic function. Our results show that dopamine dramatically reduces the direct cortical input to CA1 (the perforant path input), while having little effect on the input from CA3. In order to interpret the functional consequences of this pathway-specific modulation, it is necessary to understand the function of CA1 and the role of dopaminergic input from the ventral tegmental area (VTA). In the sixth part of this paper we consider several possibilities and address the issue of how dopamine hyperfunction or NMDA hypofunction, abnormalities that may underlie schizophrenia, might lead to the symptoms of the disease. Relevant to this issue is the demonstrated role of the hippocampus in novelty detection, a function that is likely to depend on sequence recall by the hippocampus. Novelty signals are generated when reality does not match the expectations generated by sequence recall. One possible site for computing mismatch is CA1, since it receives predictions from CA3 and sensory "reality" via the perforant path. Our data suggest that disruption of this comparison would be expected under conditions of dopamine hyperfunction or NMDA hypofunction. Also relevant is the fact that the VTA, which fires in response to novelty, may both depend on hippocampal-dependent novelty detection processes and, in turn, affect hippocampal function. Through large-scale modeling that considers both the processes performed by the hippocampus and the neuromodulatory loops in which the hippocampus is embedded, it is becoming possible to generate working hypotheses that relate synaptic function and malfunction to behavior.

Synaptic plasticity: a molecular memory switch.

Recent work shows that two molecules with major roles in synaptic plasticity--CaMKII and the NMDA receptor--bind to each other. This binding activates CaMKII and triggers its autophosphorylation. In this state, it may act as a memory switch and strengthen synapses through enzymatic and structural processes.

A model of synaptic memory: a CaMKII/PP1 switch that potentiates transmission by organizing an AMPA receptor anchoring assembly.

Ca2+/calmodulin-dependent protein kinase II (CaMKII) is localized in the postsynaptic density (PSD) and is necessary for LTP induction. Much has been learned about the autophosphorylation of CaMKII and its dephosphorylation by PSD protein phosphatase-1 (PP1). Here, we show how the CaMKII/PP1 system could function as an energy-efficient, bistable switch that could be activated during LTP induction and remain active despite protein turnover. We also suggest how recently discovered binding interactions could provide a structural readout mechanism: the autophosphorylated state of CaMKII binds tightly to the NMDAR and forms, through CaMKII-actinin-actin-(4.1/SAP97) linkages, additional sites for anchoring AMPARs at synapses. The proposed model has substantial experimental support and elucidates principles by which a local protein complex could produce stable information storage and readout.

A labile component of AMPA receptor-mediated synaptic transmission is dependent on microtubule motors, actin, and N-ethylmaleimide-sensitive factor.

Glutamate receptor channels are synthesized in the cell body, are inserted into intracellular vesicles, and move to dendrites where they become incorporated into synapses. Dendrites contain abundant microtubules that have been implicated in the vesicle-mediated transport of ion channels. We have examined how the inhibition of microtubule motors affects synaptic transmission. Monoclonal antibodies that inactivate the function of dynein or kinesin were introduced into hippocampal CA1 pyramidal cells through a patch pipette. Both antibodies substantially reduced the AMPA receptor-mediated responses within 1 hr but had no effect on the NMDA receptor-mediated response. Heat-inactivated antibody or control antibodies had a much smaller effect. A component of transmission appeared to be resistant even to the combination of these inhibitors, and we therefore explored whether other agents also produce only a partial inhibition of transmission. A similar resistant component was found by using an actin inhibitor (phalloidin) or an inhibitor of NSF (N-ethylmaleimide-sensitive fusion protein)/GluR2 interaction. We then examined whether these effects were independent or occluded each other. We found that a combination of phalloidin and NSF/GluR2 inhibitor reduced the response to approximately 30% of baseline level, an effect only slightly larger than that produced by each agent alone. The addition of microtubule motor inhibitors to this combination produced no further inhibition. We conclude that there are two components of AMPA receptor-mediated transmission; one is a labile pool sensitive to NSF/GluR2 inhibitors, actin inhibitors, and microtubule motor inhibitors. A second, nonlabile pool resembles NMDA receptor channels in being nearly insensitive to any of these agents on the hour time scale of our experiments.

Is persistent activity of calcium/calmodulin-dependent kinase required for the maintenance of LTP?

Calcium/calmodulin-dependent protein kinase II (CaMKII) is concentrated in the postsynaptic density (PSD) and plays an important role in the induction of long-term potentiation (LTP). Because this kinase is persistently activated after the induction, its activity could also be important for LTP maintenance. Experimental tests of this hypothesis, however, have given conflicting results. In this paper we further explore the role of postsynaptic CaMKII in induction and maintenance of LTP. Postsynaptic application of a CaMKII inhibitor [autocamtide-3 derived peptide inhibitor (AC3-I), 2 mM] blocked LTP induction but had no detectable affect on N-methyl-D-aspartate (NMDA)-mediated synaptic transmission, indicating that the primary function of CaMKII in LTP is downstream from NMDA channel function. We next explored various methodological factors that could account for conflicting results on the effect of CaMKII inhibitors on LTP maintenance. In contrast to our previous work, we now carried out experiments at higher temperature (33 degrees C), used slices from adult animals, and induced LTP using a tetanic stimulation. However, we still found that LTP maintenance was not affected by postsynaptic application of AC3-I. Furthermore the inhibitor did not block LTP maintenance under conditions designed to enhance the Ca(2+)-dependent activity of protein phosphatases 1 and 2B (elevated Ca(2+), calmodulin, and an inhibitor of protein kinase A). We also tested the possibility that CaMKII inhibitor might not be able to affect CaMKII once it was inserted into the PSD. In whole-brain extracts, AC3-I blocked autophosphorylation of both soluble and particulate/PSD CaMKII with similar potencies although the potency of the inhibitor toward other CaMKII substrates varied. Thus we were unable to demonstrate a functional role of persistent Ca(2+)-independent CaMKII activity in LTP maintenance. Possible explanations of the data are discussed.

Gating of human theta oscillations by a working memory task.

Electrode grids on the cortical surface of epileptic patients provide a unique opportunity to observe brain activity with high temporal-spatial resolution and high signal-to-noise ratio during a cognitive task. Previous work showed that large-amplitude theta frequency oscillations occurred intermittently during a maze navigation task, but it was unclear whether theta related to the spatial or working memory components of the task. To determine whether theta occurs during a nonspatial task, we made recordings while subjects performed the Sternberg working memory task. Our results show event-related theta and reveal a new phenomenon, the cognitive "gating" of a brain oscillation: at many cortical sites, the amplitude of theta oscillations increased dramatically at the start of the trial, continued through all phases of the trial, including the delay period, and decreased sharply at the end. Gating could be seen in individual trials and varying the duration of the trial systematically varied the period of gating. These results suggest that theta oscillations could have an important role in organizing multi-item working memory.

A cGMP-gated channel subunit in Limulus photoreceptors.

The phototransduction cascade in invertebrate photoreceptors has not been fully elucidated. It has been proposed that in Limulus ventral photoreceptor cGMP is the intracellular second messenger that directly controls the gating of the light-dependent channels (Johnson et al., 1986: Bacigalupo et al., 1991). Recently, a putative cGMP-gated channel cDNA, Lcng1, has been cloned from Limulus and shown to be expressed in the brain and the ventral eye (Chen et al., 1999). In this study, we sought to more specifically localize the LCNG1 transcript and protein. In situ hybridization was used to determine whether the gene is expressed in glia or photoreceptor cells in the ventral eye. The results clearly demonstrated that Lcngl mRNA is transcribed in the ventral photoreceptors. On Western blots probed with a polyclonal antibody raised against the C-terminus of LCNGI, a 100-kDa band and an 80-kDa band was labeled in the membrane protein preparations from brain and ventral eye, respectively. The labeling of these bands was blocked by preabsorption of the antibody with the antigen, indicating the labeling specificity. Immunocytochemistry and confocal microscopy were applied to investigate the subcellular localization of this antigen. Immunolabeling was highly localized in the transducing lobes of ventral eye photoreceptors and lateral eye photoreceptors. In both cases, the labeling was associated with membrane regions specialized for phototransduction, but the exact pattern appeared to be somewhat different in the two eyes. Preabsorption of the antiserum with antigen abolished the labeling, confirming specificity. The results lend support to the hypothesis that a cGMP-gated channel is directly involved in the phototransduction process.

Inhibitors of guanylate cyclase inhibit phototransduction in limulus ventral photoreceptors.

The second messenger systems involved in the final stages of the phototransduction cascade in Limulus photoreceptors remain unclear. Excised patches of transducing membrane contain cGMP-gated channels, suggesting the involvement of cGMP in the excitation process. To further explore this possibility, we tested the effects of inhibitors and agonists of guanylate cyclase. The active site cyclase inhibitors guanosine 5'-tetraphosphate and adenosine 5'-tetraphosphate produced a reversible reduction of the response to light without affecting resting membrane properties. The cyclase inhibitor Rp-GTPalphaS produced a similar reduction, but the effect was only slightly reversible. The reduction in the response produced by these inhibitors was robust, often producing over a 95% decrease in the amplitude of the light response. Previous work had shown that an end-product cyclase inhibitor, imidodiphosphate, also inhibited the response. The consistent results with four different guanylate cyclase inhibitors strongly support the involvement of this enzyme in the phototransduction cascade. To determine whether the guanylate cyclase involved is the NO-dependent soluble form, we applied inhibitors and activators of the nitric oxide synthase/guanylate cyclase pathway such as L-N5-(1-iminoethyl) ornithine, sodium nitroprusside, and carboxy-PTIO. None of these agents had any substantial effect on phototransduction. Taken together, these results support a role for a particulate guanylate cyclase in Limulus photoreceptor excitation.

Inhibition of the cAMP pathway decreases early long-term potentiation at CA1 hippocampal synapses.

Long-term potentiation (LTP) has several different phases, and there is general agreement that the late phase of LTP requires the activation of adenylyl cyclase (AC) and cAMP-dependent protein kinase (PKA). In contrast, several studies indicate that the early LTP is not affected by interfering with the cAMP pathway. We have further tested the role of the cAMP pathway in early LTP using several types of inhibitors. Bath application of the PKA inhibitor H89 suppressed the early LTP induced by a single tetanus. Similarly, the LTP induced by a pairing protocol was decreased by postsynaptic intracellular perfusion of the peptide PKA inhibitor PKI(6-22) amide. The decrease of LTP produced by these inhibitors was evident immediately after induction. These results indicate that PKA is important in early LTP, that its locus of action is postsynaptic, and that it does not act merely by enhancing the depolarization required for LTP induction. The failure of some other inhibitors of the cAMP pathway to affect the early phase of LTP might be attributable to the saturation of some step in the cAMP pathway during a tetanus. In agreement with this hypothesis we found that application of the AC inhibitor SQ 22536 by itself did not affect the early phase of LTP, but did produce a reduction if the cAMP pathway was already attenuated by the PKA inhibitor H89. Our analysis of the results of genetic modifications of the cAMP pathway, especially the work on AC knock-outs, indicates that the genetic data are generally consistent with the pharmacological results showing the importance of this pathway in early LTP.

Position reconstruction from an ensemble of hippocampal place cells: contribution of theta phase coding.

Previous analysis of the firing of individual rat hippocampal place cells has shown that their firing rate increases when they enter a place field and that their phase of firing relative to the ongoing theta oscillation (7-12 Hz) varies systematically as the rat traverses the place field, a phenomenon termed the theta phase precession. To study the relative contribution of phased-coded and rate-coded information, we reconstructed the animal's position on a linear track using spikes recorded simultaneously from 38 hippocampal neurons. Two previous studies of this kind found no evidence that phase information substantially improves reconstruction accuracy. We have found that reconstruction is improved provided epochs with large, systematic errors are first excluded. With this condition, use of both phase and rate information improves the reconstruction accuracy by >43% as compared with the use of rate information alone. Furthermore, it becomes possible to predict the rat's position on a 204-cm track with very high accuracy (error of <3 cm). The best reconstructions were obtained with more than three phase divisions per theta cycle. These results strengthen the hypothesis that information in rat hippocampal place cells is encoded by the phase of theta at which cells fire.

A role of actin filament in synaptic transmission and long-term potentiation.

The role of actin filaments in synaptic function has been studied in the CA1 region of the rat hippocampal slice. Bath application (2 hr) of the actin polymerization inhibitor latrunculin B did not substantially affect the shape of dendrites or spines. However, this and other drugs that affect actin did affect synaptic function. Bath-applied latrunculin B reduced the synaptic response. Several lines of evidence indicate that a component of this effect is presynaptic. To specifically test for a postsynaptic role for actin, latrunculin B or phalloidin, an actin filament stabilizer, was perfused into the postsynaptic neuron. The magnitude of long-term potentiation (LTP) was decreased at times when baseline transmission was not yet affected. Longer applications produced a decrease in baseline AMPA receptor (AMPAR)-mediated transmission. The magnitude of the NMDA receptor-mediated transmission was unaffected, indicating a specific effect on the AMPAR. These results suggest that postsynaptic actin filaments are involved in a dynamic process required to maintain AMPAR-mediated transmission and to enhance it during LTP.

Dopamine selectively inhibits the direct cortical pathway to the CA1 hippocampal region.

The perforant path input (pp) is a major direct source of specific sensory information for the CA1 hippocampal region. The termination area of this pathway, the stratum lacunosum-moleculare, has the highest concentration of dopamine receptors in the hippocampus. We have examined the properties of the pp input and its modulation by dopamine. The input is glutamatergic and has a larger NMDA component than the Schaffer collateral (sc) input. Dopamine strongly inhibits the response to pp stimulation (IC50 approximately 3 microM) but not the response to sc stimulation. Dopamine reduces both the NMDA and AMPA components of transmission at the pp and increases paired-pulse facilitation. In the sc, the NMDA component but not the AMPA component is decreased, and paired-pulse facilitation is not affected. The effect of dopamine on the pp does not depend on GABAA inhibition but is reduced by the antagonists of both D1 and D2 families of dopamine receptors. The effect is not completely blocked by the combination of D1 and D2 antagonists, but is completely blocked by the atypical neuroleptic clozapine. Our results provide the first evidence for strong dopaminergic control of transmission in the perforant path. By inhibiting this pathway, dopamine hyperfunction and/or NMDA hypofunction abnormalities implicated in schizophrenia may isolate CA1 from its main source of sensory information.

An oscillatory short-term memory buffer model can account for data on the Sternberg task.

A limited number (7 +/- 2) of items can be held in human short-term memory (STM). We have previously suggested that observed dual (theta and gamma) oscillations could underlie a multiplexing mechanism that enables a single network to actively store up to seven memories. Here we have asked whether models of this kind can account for the data on the Sternberg task, the most quantitative measurements of memory search available. We have found several variants of the oscillatory search model that account for the quantitative dependence of the reaction time distribution on the number of items (S) held in STM. The models differ on the issues of (1) whether theta frequency varies with S and (2) whether the phase of ongoing oscillations is reset by the probe. Using these models the frequencies of dual oscillations can be derived from psychophysical data. The derived values (ftheta = 6-10 Hz; fgamma = 45-60 Hz) are in reasonable agreement with experimental values. The exhaustive nature of the serial search that has been inferred from psychophysical measurements can be plausibly explained by these oscillatory models. One argument against exhaustive serial search has been the existence of serial position effects. We find that these effects can be explained by short-term repetition priming in the context of serial scanning models. Our results strengthen the case for serial processing and point to experiments that discriminate between variants of the serial scanning process.