If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Stimulus-Selective Response Plasticity in the Visual Cortex: An Assay for the Assessment of Pathophysiology and Treatment of Cognitive Impairment Associated with Psychiatric Disorders
Address correspondence to Mark Bear, Ph.D., Massachusetts Institute of Technology, The Picower Institute for Learning and Memory, Howard Hughes Medical Institute, 43 Vassar Street, 46-3301 Cambridge, MA 02139
Long-term potentiation (LTP) is a form of experimentally induced enhancement of chemical synaptic transmission that has long been proposed as a model of the endogenous processes of synaptic plasticity that mediate memory. There is a large body of evidence that the molecular mechanisms underlying experimentally induced LTP also subserve various forms of naturally occurring, experience-dependent synaptic plasticity in animals and humans. Here we describe a phenomenon called stimulus-specific response potentiation (SRP), which occurs in the primary visual cortex of mice as a result of repeated exposure to visual stimuli and is believed to reveal the mechanisms that underlie perceptual learning. We first describe evidence that SRP represents naturally occurring LTP of thalamo-cortical synaptic transmission. We then discuss the potential value of SRP as a preclinical assay for the assessment of putative drug treatments on synaptic plasticity. Stimulus-specific response potentiation is not only easy to assay and robust but captures features of feed-forward glutamatergic function and visual learning that are deficient in human psychiatric disorders, notably including schizophrenia. We suggest that phenomena analogous to SRP in humans are likely to be useful biomarkers of altered cortical LTP and of treatment response in diseases associated with impaired cognition.
). Symptoms of schizophrenia are separated into major clusters. These include positive, negative, and cognitive symptoms. Early descriptions of schizophrenia, initially designated dementia praecox, viewed the cognitive elements of the disorder as a central feature (
). Although available treatments have focused on positive symptoms or psychosis, there is now increased emphasis on tackling cognitive deficits in schizophrenia (
). Although it has long been held that schizophrenia results from functional disruption of particular brain regions, such as the prefrontal cortex and the hippocampus (
). Thus, overall pathophysiology of cognitive dysfunction cannot be attributed to a specific locus. Instead, deficits across the brain, including early in sensory processing, are likely to contribute different elements to the disorder.
Deficits in visual learning have been observed in schizophrenic subjects (
) but, regardless, could be exploited to reveal pathophysiological processes at the heart of the cognitive deficit. Human perceptual learning can be highly specific for features such as spatial frequency or orientation, suggesting that primary sensory cortex—in which these primitive features have not yet been integrated into invariant representations (
). Therefore understanding the synaptic basis of experience-dependent plasticity in primary visual cortex (V1) might provide insight into schizophrenia pathophysiology.
Here we describe a simple, easy-to-assay form of stimulus-selective response potentiation (SRP) in the rodent that is well-understood at the molecular level. Stimulus-specific response potentiation is a manifestation, in part, of synaptic modification at the thalamocortical synapses in V1 and uses the mechanisms of canonical long-term potentiation (LTP). We suggest that the SRP assay in mice will enable interrogation of physiological mechanisms that might go awry in schizophrenia and other mental illnesses characterized by impaired cognitive function. Moreover, because it is a natural and functionally relevant manifestation of LTP, it enables assessment of the in vivo efficacy of potential treatments that target this mechanism.
SRP
Stimulus-specific response potentiation is long-lasting, experience-dependent plasticity that occurs in V1 of mice (Figure 1). In awake head-fixed animals, averaged visual evoked potentials (VEPs) driven by phase-reversing sinusoidal grating stimuli, recorded from layer 4 of binocular V1, undergo a gradual but striking increase in amplitude across days of repeated presentation of the same stimulus (
). Once expressed, SRP lasts for at least several weeks, and no evidence has yet been produced that it degrades with time, even in the absence of continued stimulus presentation. Alteration of stimulus properties, typically through changing orientation, reveals the specificity of the effect. Even stimuli rotated by as little as 5° from the familiar stimulus drive significantly lower amplitude VEPs (Figure 2A). Altering the spatial frequency (Figure 2B) or the contrast of the stimulus (Figure 2C) has a similar effect. Most remarkably, if the stimulus is presented to one eye only, SRP can be induced through this experienced eye without transfer to the inexperienced eye (Figures 2D–2F). This input specificity strongly suggests that plasticity underlying SRP occurs within layer 4, at thalamo-cortical synapses. At this stage of visual processing in the mouse visual cortex, eye-specific thalamic inputs converge on the same population of layer 4 target cells in the binocular zone (Figure 2E). This input specificity is reminiscent of LTP.
Figure 1Stimulus-specific response potentiation (SRP). (A) Visual-evoked potentials (VEPs) are recorded from both hemispheres of head-fixed, awake, male C57BL/6 mice and driven by full-field, high-contrast, phase-reversing sinusoidal grating stimuli. Electrodes are positioned in layer 4 of the binocular region of primary visual cortex (V1). (B) Diagram describing the time-course of a typical SRP experiment. Each block represents a day. Under anesthesia, animals receive implantation bilaterally with recording electrodes, and a headpost is affixed. They are then allowed to recover overnight, before receiving 2 half-hour habituation sessions in front of a gray screen during the following 2 days. On the next day, experimental day 1, a first oriented stimulus is presented (X°), and an averaged VEP resulting from 400 phase reversals is generated. This procedure is then repeated over the next 4 experimental days. On experimental day 5 this now-familiar stimulus is presented and, interleaved, a novel X+45° stimulus is also presented. This novel stimulus is then presented over the following 4 days, and VEPs are collected before finally testing VEP amplitude driven by 3 stimuli, the two familiar X and X+45° orientations and a novel X+90° stimulus. (C) The phenomenon of SRP can be seen in the gradual enhancement in VEP amplitude across days. The SRP develops over 5 days to X°, and orientation specificity is revealed as the VEP falls back to basal amplitude when the X+45° stimulus is presented. The SRP then develops with a similar temporal profile to the novel stimulus. Returning to the original X° stimulus, even without presentation during the intervening 4 days, reveals maintained VEP amplitude potentiation. Stimulus specificity continues to be apparent with each presentation of a novel stimulus as the VEP falls back to basal amplitude. Scale bars are 100 μV vertically and 50 μsec horizontally (data re-presented from [
Figure 2SRP specificity. (A) The SRP is exquisitely specific for orientation, as revealed by the fact that, on day 5 after saturated SRP to a now-familiar orientation (dark gray square), VEP amplitude drops significantly back toward basal values with a novel (Nov) stimulus (dark gray circles) shifted by just 5° from the familiar (Fam) orientation. If the shift in orientation is gradually increased, it can be seen that a novel stimulus shifted by approximately 25°–30° from familiar will evoke a VEP of basal amplitude. (B) The SRP is also specific for spatial frequency in that repeated binocular exposure to a .05 cycle/° grating of a particular orientation (dark gray squares) results in SRP only to that spatial frequency with no significant gains over a novel orientation (dark gray circles) at other spatial frequencies. (C) The same is true for contrast. The SRP at 100% contrast shows no transfer to other contrasts (data re-presented from [
], with permission from Nature Publishing Group). (D) An occluding paddle can be used to selectively block the presentation of the visual stimulus to one or the other eye of the headfixed mouse preparation and thereby test VEP amplitudes selectively evoked with binocular (dark gray), Contralateral (black), or Ipsilateral (light gray)-only stimulus presentation. (E) In the rodent visual cortex, thalamo-recipient cells—predominantly in layer 4 but also found in other layers—are usually responsive to independent inputs from either eye. This feature of the system enables testing of whether SRP transfers from an experienced eye to an inexperienced eye. (F) The SRP is revealed to be eye-specific. Presentation of two different stimuli (circle or square) to only the Ipsilateral (light gray) or only the Contralateral eye (black) on experimental day 1 allows for determination of basal VEP amplitude. In this case only the Ipsilateral eye then experiences repeated presentation of the same stimulus over subsequent days until a final test day when stimulus-specificity is tested by reversing the configuration of stimulus presentation to each eye that was seen on day 1, such that both eyes view a novel stimulus. There is no transfer of SRP from the Ipsilateral to the Contralateral eye (data re-presented from [
Long-term potentiation is a phenomenon in which chemical synaptic transmission is enhanced in a lasting fashion through co-incident pre- and post-synaptic activity. It has several properties that make it an attractive candidate memory mechanism. First, LTP has been demonstrated to last for several months under the right conditions, revealing that it has the longevity required of a storage device (
). Second, it is input-specific: only active synapses contacting a depolarizing cell undergo LTP, whereas nearby inactive neighboring synapses remain unchanged in strength (Figure 3). Third, it is both associative and cooperative (
). Thus, weakly active synapses, themselves not strong enough to drive post-synaptic depolarization, can undergo potentiation if associated in time with activity in other inputs that cooperate to depolarize the post-synaptic target cell (
) when describing the theoretical constraints of a synaptic memory mechanism in the brain, leading to comparable forms of plasticity commonly being described as “Hebbian.”
Figure 3Canonical long-term potentiation (LTP). (A) The transverse hippocampal slice has been the favored preparation for studying LTP, due to its beautifully simple laminated structure, the presence of a single layer of excitatory cells organized as a perfect dipole, and its overall robustness ex vivo. The synapse between Schaffer collateral (Sch) fibers originating from pyramidal cells in the CA3 subfield, either ipsilaterally or contralaterally via commissural fibers (comm.), and pyramidal cells in the CA1 subfield has served as a primary focus of LTP research; and the molecular mechanisms of induction, expression, and maintenance are now described as canonical. Very different mechanisms are thought to underlie mossy fiber (mf)-CA3 LTP and will not be discussed here. The LTP at synapses between the perforant path (pp), originating in the entorhinal cortex (EC), and granule cells of the dentate gyrus (DG) were the first to be studied, and the mechanisms are broadly similar to the canonical LTP described here. Visualization of the slice allows for placement of two stimulating electrodes (Stim.) in pathways that independently contact the same population of post-synaptic pyramidal cells—here described as input 1 and 2. (B) This organization has enabled researchers to demonstrate the input-specificity of LTP, here revealed as enhancement of the excitatory postsynaptic potential (EPSP) slope of population field potentials recorded in the dendritic zone of CA1, selectively in the tetanized input (black circles), which receives high-frequency stimulation (HFS) without impact on the untetanized control pathway (open circles), which receives just low-frequency stimulation. The LTP can be maintained over many hours if the appropriate induction protocol is used. Scale bar is 1 mV vertically and 5 msec horizontally. (C) Similar molecular mechanisms are believed to support LTP at TC synapses. Again, the ex vivo slice preparation has been critical in identifying molecular mechanisms, although in this case the structure is primary visual cortex. Several synapses can be studied in this preparation, but here we focus on the TC input from the thalamus to layer 4 and layer 2/3. Stimulation of the white matter containing TC fibers evokes population field potential responses recorded in layer 4 and layer 2/3. (D) The LTP can be induced at these synapses with θ burst stimulation (TBS), resulting in a potentiation of the amplitude of the field potential. Field potential amplitude rather than EPSP slope is used as a proxy for synaptic potentiation, because pure dendritic field potentials cannot be isolated in this multi-laminar structure. Scale bar is 1 mV vertically and 10 msec horizontally (plot adapted from [
], with permission from Nature Publishing Group). (E) Canonical LTP mechanisms center on the N-methyl-D-aspartate receptor (NMDAR) and the influx of calcium (Ca2+) through that receptor as a result of both pre-synaptic glutamate release and post-synaptic depolarization, predominantly mediated by α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid class of glutamate receptor (AMPAR) conductance of cations. The Ca2+ influx initiates kinase and phosphatase activity, the balance of which determines whether LTP is expressed. Expression mechanisms include post-translational modification (phosphorylation) of existing AMPARs by calcium-dependent kinases to enhance conductance and the introduction of new AMPARs into the post-synaptic membrane. There also exist pre-synaptic expression mechanisms that lead to an enhancement of glutamate release. Pre-synaptic expression mechanisms require the existence of a retrograde signal for which several candidates have been identified. For maintenance of LTP beyond a few hours there is a requirement for the synthesis of new proteins, through both local translation and transcription of RNA at the nucleus. Again, a signal from the synapse to the nucleus is required for this de novo synthesis to occur, and evidence suggests that a signaling cascade is initiated by sustained calcium influx that activates kinases such as CaMKII, PKA, and PKC. The resultant newly expressed LTP effector proteins are believed to participate in further stabilization of new AMPAR insertion and structural plasticity, resulting in growth of the synapse both pre- and post-synaptically. Mg2+, magnesium dimer; Na+, sodium cation.
Long-term potentiation was initially induced by applying a high-frequency train of electrical pulses or “tetanus” to an afferent pathway in the hippocampus (
). This tetanus ensured coincident pre- and postsynaptic depolarization and resulted in a lasting enhancement of synaptic strength, as revealed by measuring the slope of the excitatory postsynaptic potential (EPSP) component of electrically evoked population field potentials. The same phenomenon has since been observed in awake, freely moving animals and monitored for many days, weeks, and even months (
), an ex vivo preparation that allows for precisely timed pharmacological treatment, has enabled a careful dissection of the biochemical/biophysical mechanisms of LTP. More recently, genetic manipulations have augmented this approach by increasing the specificity and degree of intervention. With the Schaffer collateral-CA1 pyramidal cell synapse as a model, mechanisms of LTP-induction, expression, and maintenance have been determined (Figure 3E). Many of these mechanisms are shared across synapses in the nervous system, including in V1 (Figures 3C and 3D), and can be described as canonical, although it is important to remember that characteristics of LTP do vary regionally (
). Recent work with genetically modified animals and in vivo pharmacology has demonstrated commonalities between molecular mechanisms of LTP and memory (see [
Long-term potentiation-like effects can also be induced in the brain through sensory experience. In the hippocampus, synaptic potentiation develops concurrently with two forms of memory; trace eyeblink conditioning (
), demonstrating shared core mechanisms. These experiments have all been important in establishing a causal relationship between LTP-like plasticity and memory in the central nervous system, although it has still not been proven that synaptic potentiation is both necessary and sufficient for memory (
) or, most importantly, exactly how synaptic plasticity is implemented at a circuit and systems level.
Is SRP an Endogenous Form of LTP?
Evidence for a causal relationship between LTP and SRP comes from in vivo experiments in which thalamo-cortical LTP is induced by delivery of an electrical tetanus to the dorsal lateral geniculate nucleus. This form of LTP, as initially assessed with electrically evoked field potentials, also enhances the amplitude of VEPs in anesthetized rats (
), indicating that electrical stimulation can have a lasting potentiating influence on sensory processing. This observation has since been replicated by others in the visual cortex (
NR2B subunit-dependent long-term potentiation enhancement in the rat cortical auditory system in vivo following masking of patterned auditory input by white noise exposure during early postnatal life.
). Similar LTP can be induced in the mouse, and it has been possible to use the principles of mimicry and occlusion to test the effects of LTP on the subsequent induction of SRP (
) (Figure 4). In the mouse, just as in the rat, the amplitude of both the electrically evoked field potential and the VEP is significantly enhanced by the induction of thalamo-cortical LTP in a single hemisphere. The LTP occludes the subsequent development of SRP in this hemisphere, whereas SRP develops normally in the opposite control hemisphere that has not undergone LTP (Figure 4C). The occlusion of SRP by LTP is mutual, as demonstrated by inducing SRP to a single orientation before LTP. The VEPs driven by this familiar stimulus do not undergo LTP. Due to the stimulus specificity of SRP, it is possible to also show a novel stimulus both before and after LTP and demonstrate, in the same animals and same recording sessions, that VEPs are potentiated by LTP only if the stimulus is novel (Figure 4D). Thus, all the available evidence, both correlative and interventional, indicates that SRP and LTP use the same processes and that SRP can be regarded as a proxy for N-methyl-D-aspartate receptor (NMDAR)-dependent feedforward excitatory synaptic plasticity.
Figure 4SRP as LTP. (A) LTP can be induced at thalamo-cortical synapses with an acutely placed stimulating electrode in the dorsal lateral geniculate nucleus (dLGN) under light isoflurane anesthesia. Field potentials can be recorded in layer 4 of V1 through the same recording electrode, whether driven by an electrical pulse to the dLGN in the anesthetized mouse or by a phase reversal of a visual stimulus in the awake animal. Experiments could then be conducted to assess interactions between SRP and LTP with one hemisphere undergoing LTP (black) and the opposite hemisphere serving as a control receiving a single electrical pulse every 30 sec (white). (B) Electrically evoked TC field potentials recorded in layer 4 underwent significant LTP after three trains of TBS were applied with a 5-min inter-burst interval. The LTP of field potential amplitude lasted for at least 1 hour before the stimulating electrode was removed and the animal recovered from anesthesia. Scale bar is 1 mV vertically and 5 msec horizontally. (C) The LTP mimics SRP in enhancing VEP amplitude in the tetanized hemisphere (black circles) significantly compared with VEPs in the control hemisphere (white circles) immediately upon recovery from anesthesia. The SRP is also significantly occluded because, by day 5, VEPs are of equivalent amplitude in both hemispheres, with those in the untetanized hemisphere undergoing significantly more SRP than in the tetanized hemisphere. (D) The complementary experiment reveals that the occlusion effect is bidirectional, such that TBS delivery enhances VEP amplitude in a separate group of mice after SRP has already been saturated. However, the mimicry effect is only seen for VEPs driven by novel stimuli (black circles) and not for those driven by familiar stimuli (black squares), due to occlusion by prior SRP. Thus VEPs driven by a familiar stimulus are significantly greater amplitude than those driven by a novel stimulus before TBS delivery but no longer different after TBS delivery. Scale bar is 100 μV vertically and 50 msec horizontally (adapted from [
We hasten to add, however, that VEP amplitude undergoes much greater potentiation through experience or electrical stimulation than electrically evoked potentials do during in vivo LTP experiments. Thus, the effects of simple Hebbian synaptic plasticity are amplified when the VEP is generated by a visual stimulus. Further experiments will be required to fully understand this amplification process. It is interesting to consider the possibility that inhibitory interneurons might play a significant role. In any case, although we can conclude that SRP uses the mechanisms of LTP, the mechanisms of LTP are not sufficient to account for the full expression of SRP.
The behavioral significance of SRP has not yet been established. In many ways the phenomenon is reminiscent of perceptual learning, a long-studied form of plasticity that results in improved stimulus detection (
), although in these cases learning was accompanied by the reward of finding an escape from water. Thus, perceptual learning seems to enhance the strength of V1 synapses, although results are somewhat inconsistent. One study reports a subsequent occlusion of LTP at intra-cortical synapses (
). These contrasting findings reveal the added complexity of including explicit reward in the experience of the animal. It is known, for example, that synaptic plasticity can occur in V1 as a result of reward (
There are several reports in both animals and humans that LTP-like plasticity can be induced in V1 with a photic tetanus, in which flashing or phase reversing visual stimuli are presented at a frequency sufficiently high to induce an immediate enhancement of VEP amplitude (
) will discuss this phenomenon at greater length in an accompanying article within this issue. Here they reveal, importantly, that photic-tetanus-induced potentiation is accompanied by an improvement in detection of stimuli at low luminance, demonstrating that this form of plasticity has some behavioral utility. We would like to point out that there are both similarities and differences between SRP and this tetanus-induced LTP. The most striking similarities are, first, that molecular mechanisms seem to be shared (
). In the future, it will be interesting to test whether these two forms of plasticity mutually occlude one another. An obvious difference that exists between SRP and the effects of tetanic stimulation, either photic or electrical, is that SRP takes several hours to impact VEP amplitude, whereas tetanus-induced LTP expresses itself almost immediately. This delayed manifestation of SRP suggests a requirement for some kind of consolidation process, potentially involving sleep (
). It will be important to characterize this process. It is possible that behavioral modifications accompanying SRP might be delayed in a similar fashion, in keeping with a causal role for the plasticity in visual learning.
Molecular Mechanisms of LTP and SRP
We will not discuss the molecular mechanisms of canonical LTP in great depth, because it is beyond the scope of this review, but we will introduce important features that are shared with SRP.
The detection of coincident activity pre- and post-synaptically—such a key feature of canonical Hebbian LTP—requires the NMDA class of glutamate receptor (
). The NMDAR is both ionotropic and voltage-dependent. This means that glutamate must be present to open the receptor ion channel, indicative of pre-synaptic activity and transmitter release, and the post-synaptic membrane must also be depolarized to remove a blocking magnesium ion from the channel (
). If both of these events occur, then there is a large cation influx through the receptor channel, notably of calcium, and the induction mechanisms of LTP are initiated (
). Several calcium-dependent kinases are then activated in response to the elevated levels of calcium and phosphorylate target substrates within the synapse that participate in the expression of LTP (
). The α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid class of glutamate receptor (AMPAR), which mediates the majority of fast excitatory transmission in the central nervous system, is notable amongst these targets. Phosphorylation of the receptor at key residues alters its conformation and enhances its function (
). In its early phase, LTP does not require the synthesis of new proteins. However, after an hour or so there is a requirement for new proteins at the synapse if LTP is to be maintained (
). This enzyme is unusual in that it has no auto-regulatory domain and does not require elevated calcium to remain active. It is believed to maintain LTP for long periods of time by preventing AMPAR removal from the post-synaptic membrane (
). There are a wide range of other LTP induction and expression mechanisms, including pre-synaptic effects that increase transmitter release and changes in the structure of the synapse, but we do not provide any detail about those here, because it is not within the scope of this review (for further reading see [
Long-term potentiation in the dentate gyrus of the anaesthetized rat is accompanied by an increase in extracellular glutamate: Real-time measurements using a novel dialysis electrode.
Philos Trans R Soc Lond B Biol Sci.2003; 358: 675-687
The results of experimental interventions strongly suggest that SRP represents an LTP-like process. Systemic application of the competitive NMDAR antagonist 3-(2-Carboxypiperazin-4-yl)propyl-1-phosphonic acid (CPP) before each stimulus-exposure prevents the development of SRP, just as has been observed for visual cortical LTP (
) (Figures 5A and 5B). Importantly, genetic ablation of the NMDAR selectively in glutamatergic cells of the cortex also prevents the induction of SRP, demonstrating that SRP results from plasticity at synapses on these excitatory neurons (
). Local blockade of AMPAR insertion with a peptide that mimics the cytoplasmic tail of AMPARs and thereby exerts a dominant-negative effect on insertion prevents SRP (
), restricted to the recording site, can prevent SRP induction or maintenance, respectively. Thus, plasticity subserving SRP occurs in V1 and is not an echo of events occurring elsewhere in the brain. The VEP comprises synaptic events (
), so SRP likely represents LTP-like plasticity of thalamo-cortical synapses in layer 4. Nevertheless, there is evidence that intra-cortical plasticity occurs through perceptual learning (
), so we currently maintain an open mind on the degree to which plasticity mediating SRP is distributed in V1.
Figure 5Molecular mechanisms of SRP. (A) Schematic of a glutamatergic synapse, known molecular players in SRP, and interventional approaches to blocking them. (B) Repeated systemic injections of the NMDAR antagonist 3-(2-Carboxypiperazin-4-yl)propyl-1-phosphonic acid (CPP) (open symbols) before stimulus exposure prevent mice from undergoing SRP, unlike control saline injections (black symbols), demonstrating a role for the NMDAR in SRP induction. Binocular VEP data are normalized to the amplitude of the Ipsi-only response. (C) Local viral-mediated expression in V1 of a peptide that mimics the cytoplasmic tail of the GluR1 AMPAR subunit (GluR1-CT, open circles) and thereby acts in a dominant-negative fashion to prevent AMPAR insertion, also blocks SRP from developing, whereas a green fluorescent protein (GFP) control infection does not (black circles). This experiment provides evidence both that SRP requires AMPAR insertion for expression and that plasticity within V1 local to the recording site mediates SRP (adapted from [
]). (D) After saturated SRP induction, local delivery of a specific inhibitor peptide (ZIP) of the calcium-dependent protein kinase (PKC) isoform, PKMζ, erases already induced SRP such that familiar (F) (gray circle) and novel (N) stimuli (gray square) evoke VEPs of equivalent basal magnitude where a clear SRP effect had been seen before application (adapted from [
], with permission from Nature Publishing Group). (E) Mutant mice that do not express the activity-dependent protein Arc/Arg3.1 (open symbols) fail to exhibit SRP. Their wildtype littermates (black symbols) undergo normal SRP. It is, as yet, unclear how Arc/Arg3.1 might impact synaptic potentiation to prevent SRP induction (adapted from [
) (Figures 5A and 5E). This cytoskeletal-associated protein has been of great interest in the field of memory and synaptic plasticity, because its expression is activity-dependent and driven by both learning (
Inhibition of activity-dependent arc protein expression in the rat hippocampus impairs the maintenance of long-term potentiation and the consolidation of long-term memory.
). It is unclear, however, how Arc/Arg3.1 disrupts LTP mechanistically. Arc/Arg3.1 has been shown to play a clear role in the removal of AMPARs from the synaptic membrane (
), central to the process of long-term depression but of no clear relevance to LTP itself. The mechanistic role that Arc/Arg3.1 plays in SRP is, therefore, yet to be determined.
Utility of SRP in Psychiatric Research
Typical neuroleptic drugs, commonly used to treat schizophrenia, target the dopaminergic neurotransmitter system. These substances, such as chlorpromazine and haloperidol, treat positive symptoms but have little impact on cognitive symptoms (
). Psychotomimetic substances that impact dopaminergic or serotonergic systems—such as amphetamine or lysergic acid diethylamide, respectively—selectively mimic the positive symptoms of schizophrenia without major cognitive effect (
Comparative and interactive human psychopharmacologic effects of ketamine and amphetamine: Implications for glutamatergic and dopaminergic model psychoses and cognitive function.
). Thus, there is a strong argument that cognitive symptoms do not result from the same imbalance in monoaminergic transmission that is implicated in the positive symptoms. Psychotomimetic NMDAR antagonists such as phencyclidine (PCP) and ketamine do mimic the full range of schizophrenia symptoms, including cognitive symptoms, in normal individuals (
). This observation has given rise to the glutamatergic hypothesis, which asserts that alterations of NMDAR-mediated synaptic transmission/plasticity are causal dysfunctions in schizophrenia (
Severe impairment of NMDA receptor function in mice carrying targeted point mutations in the glycine binding site results in drug-resistant nonhabituating hyperactivity.
), which have behavioral manifestations similar to normal mice under the influence of PCP.
Cognitive symptoms of schizophrenia include deficits in attention, working memory, and executive function, but there is now also considerable evidence for sensory deficits as assessed, for example, with pre-pulse inhibition (
). The NMDAR might play a crucial role in the uniquely nonlinear character of magnocellular responses that allows for an amplification of cortical activity driven by visual stimulation in this pathway (
). There are also two other ways in which a reduction of NMDAR function could impact sensation. First, LTP-like plasticity necessary for perceptual learning might be disrupted. Long-term potentiation-like effects, induced noninvasively in the cortex of human subjects with repetitive transcranial magnetic stimulation, are reduced in schizophrenic patients (
). Also, accompanying articles in this issue of Biological Psychiatry describe deficits in sensory tetanus-induced event-related potential potentiation in schizophrenic subjects (
). The psychotomimetic actions of noncompetitive NMDAR antagonists such as PCP and ketamine are also believed to arise, in large degree, from blockade of NMDAR expressed in these inhibitory cells (
). Despite all of the evidence that we have presented in this article revealing that SRP relies upon feedforward Hebbian plasticity in excitatory cells of the cortex, we cannot exclude the possibility that additional mechanisms are involved. Indeed, as we have already discussed, there seems to be some amplification of feedforward Hebbian plasticity in the dramatic enhancement of VEP amplitude as a stimulus becomes familiar during SRP. We have not directly investigated this process, but an obvious hypothesis is that modifications of inhibition act synergistically with plasticity in excitatory cells to discriminate novel and familiar stimuli. At the moment this is pure speculation. Further experiments will be required to establish whether inhibition plays a role in SRP and, if so, whether the process is also NMDAR-dependent.
Although this review has very clearly made the argument that sensory experience induces plasticity within primary sensory cortex, we would be remiss to ignore the fact that top-down influences mediating attention can also have a major impact on sensory-evoked potentials (
). Therefore, we cannot exclude the possibility that, just as with the potential contribution of inhibition discussed in the preceding text, top-down effects mediate some of the transformative effects of repeated visual experience on the VEP and the failure of these effects in schizophrenia due to dopamine dysregulation. Nonetheless, it is clear from the work presented here that plasticity intrinsic to visual cortex is a necessary component for the stimulus-specific potentiation of VEPs through repeated visual presentation.
Although schizophrenia has been the focus of this article, it is important to note that alterations in visual cortical plasticity have been observed in other psychiatric disorders, such as depression (
). Stimulus-specific response potentiation might, therefore, serve as a generalized pre-clinical assay of cortical plasticity. It also seems likely, given the molecular similarities that exist in plasticity across cortical regions, that other sensory modalities could be evaluated in the same fashion. For instance, an auditory tetanus, comprising trains of tone pips, has been shown to potentiate the amplitude of auditory-evoked potentials in humans (
). This use of direct sensory stimulation to induce plasticity has a number of advantages over other pre-clinical assays, such as electrically induced LTP: First, it is noninvasive and not obviously aversive. Second, it is relatively naturalistic, and finally, it is beautifully stimulus-specific, allowing for repeated testing across time within the same individual. For these reasons sensory-induced plasticity could provide an assay of normal, dysfunctional, and drug-treated cortical plasticity that is readily translatable from animals to humans.
Conclusions
Here we have described the phenomenon of SRP and all of the evidence that strongly implicates it as a manifestation of Hebbian synaptic plasticity. There is still much to be understood about the underlying mechanisms of SRP, and some mystery surrounds its behavioral relevance. We have also yet to determine its generality across species, although similar effects in rats and humans are described in accompanying articles in this issue of Biological Psychiatry. Nevertheless, SRP is a robust and easy-to-assay form of NMDAR-dependent cortical plasticity that occurs in the intact awake mouse. Because it uses the same mechanisms as LTP and shares many properties with sensory perceptual learning, SRP should be valuable for understanding the pathophysiology of cognitive impairment in schizophrenia and for evaluation of potential therapies.
The primary research discussed in this review article was supported by grant funding from the National Institute of Mental Health , the National Eye Institute , the Howard Hughes Medical Institute , and the Picower Institute Innovations Fund .
We thank Dr. Idil Çavuş and Dr. Dan Mathalon for their helpful comments on the contents of this article during its preparation.
Dr. Cooke has no biomedical financial interests or potential conflicts of interest to report. Dr. Bear has a financial interest in Seaside Therapeutics.
References
Saha S.
Chant D.
Welham J.
McGrath J.
A systematic review of the prevalence of schizophrenia.
NR2B subunit-dependent long-term potentiation enhancement in the rat cortical auditory system in vivo following masking of patterned auditory input by white noise exposure during early postnatal life.
Long-term potentiation in the dentate gyrus of the anaesthetized rat is accompanied by an increase in extracellular glutamate: Real-time measurements using a novel dialysis electrode.
Philos Trans R Soc Lond B Biol Sci.2003; 358: 675-687
Inhibition of activity-dependent arc protein expression in the rat hippocampus impairs the maintenance of long-term potentiation and the consolidation of long-term memory.
Comparative and interactive human psychopharmacologic effects of ketamine and amphetamine: Implications for glutamatergic and dopaminergic model psychoses and cognitive function.
Severe impairment of NMDA receptor function in mice carrying targeted point mutations in the glycine binding site results in drug-resistant nonhabituating hyperactivity.
00900-0&id=gr1.jpg){kind=link}
00900-0&id=gr2.jpg){kind=link}
00900-0&id=gr3.jpg){kind=link}
00900-0&id=gr4.jpg){kind=link}
00900-0&id=gr5.jpg){kind=link}