Stimulation of dopamine (DA) receptors in the striatum is essential for voluntary motor activity and for the generation of plasticity at corticostriatal synapses. In the present study, mice lacking DA D1 receptors have been used to investigate the involvement of the D1-like class (D1 and D5) of DA receptors in locomotion and corticostriatal long-term depression (LTD) and long-term potentiation (LTP). Our results suggest that D1 and D5 receptors exert distinct actions on both activity-dependent synaptic plasticity and spontaneous motor activity. Accordingly, the ablation of D1 receptors disrupted corticostriatal LTP, whereas pharmacological blockade of D5 receptors prevented LTD. On the other side, genetic ablation of D1 receptors increased locomotor activity, whereas the D1/D5 receptor antagonist SCH 23390 decreased motor activity in both control mice and mice lacking D1 receptors.
Endogenous DA stimulated D1 and D5 receptors in distinct subtypes of striatal neurons to induce, respectively, LTP and LTD. In control mice, in fact, LTP was blocked by inhibiting the D1–protein kinase A pathway in the recorded spiny neuron, whereas the striatal nitric oxide-producing interneuron was presumably the neuronal subtype stimulated by D5 receptors during the induction phase of LTD.
Understanding the role of DA receptors in striatal function is essential to gain insights into the neural bases of critical brain functions and of dramatic pathological conditions such as Parkinson's disease, schizophrenia, and drug addiction.
- basal ganglia
- in vitro electrophysiology
- long-term depression
- long-term potentiation
- nitric oxide
Dopamine (DA) signaling in the striatum plays a central role in a variety of motor and cognitive activities. Abnormal striatal DAergic transmission is involved in several neuropsychiatric diseases, such as parkinsonism, schizophrenia, and drug addiction (Berke and Hyman, 2000; Lewis and Lieberman, 2000; Obeso et al., 2000). Endogenous DA, released from midbrain DA neurons, modulates striatal function by interacting with DA receptors. Among the various subtypes of DA receptors, the D1-like family has been involved in the regulation of motor activity and in the expression of activity-dependent synaptic plasticity at corticostriatal synapses. Accordingly, pharmacological inhibition of D1like receptors reduces spontaneous motor activity (Meyer et al., 1993; Vallone et al., 2000) and prevents both long-term depression (LTD) and long-term potentiation (LTP) (Calabresi et al., 1992a, 2000; Centonze et al., 2001; Kerr and Wickens, 2001). To date, however, it is still unknown which member of the D1-like family of DA receptors (D1 or D5) mediates these actions of DA in the striatum. In this respect, both D1 and D5 receptors are expressed in the striatum (Bergson et al., 1995; Surmeier et al., 1996; Yan and Surmeier, 1997; Rivera et al., 2002a), are positive regulators of adenylyl cyclase activity (Stoof and Kebabian, 1981; Grandy et al., 1991; Sunahara et al., 1991; Tiberi et al., 1991; Vallone et al., 2000), and might be, in principle, equally important for both motor activity and ordered synaptic plasticity. However, the evidence that the quantitative ratios of these receptors differ significantly in the various neuronal populations of the striatum (Bergson et al., 1995; Surmeier et al., 1996; Yan and Surmeier, 1997; Rivera et al., 2002a) supports the concept that they serve distinct physiological roles.
In the present study, therefore, we used mice in which the expression of DA D1 receptors was selectively disrupted to analyze the involvement of D1 and D5 receptors in locomotor activity and corticostriatal LTD and LTP.
Materials and Methods
Male wild-type (WT) and D1 DA receptor knock-out (D1-/-) mice (Xu et al., 1994) (2–3 months of age) were used for all the experiments.
Locomotor activity. For locomotor activity studies, we used a multicage activity meter system (Digiscam Animal Activity Monitor; Columbus Instruments, Columbus, OH). This apparatus consisted of eight individual mice cages (21 × 21 × 30 cm), equipped with two sets (one above the other) of eight photocell beams per side spaced 2.5 cm to measure horizontal and vertical activity. WT and D1-/- mice were habituated to the cages for 3 consecutive days, and basal activity was recorded for 3 hr on the following day. The motor-suppressing effect of SCH 23390 (a D1/D5 receptor antagonist) at the doses of 30, 50, 100, and 300 μg/kg was tested in both mouse genotypes in 3 hr sessions. Each group was composed of eight animals, and each animal was used as its own control. All injections were administered intraperitoneally, in 1 ml/100 gm body weight/volume, and SCH 23390 (Tocris Cookson, Bristol, UK) was dissolved in saline.
Electrophysiology. Intracellular and whole-cell patch-clamp electrophysiological recordings were performed in vitro from brain slices. The preparation and maintenance of coronal corticostriatal slices have been described previously (Calabresi et al., 1997, 2000; Centonze et al., 2002b). Briefly, coronal slices (200–300 μm) were prepared from tissue blocks by use of a vibratome. The slices included the neostriatum and the neocortex. A single slice was transferred to a recording chamber and submerged in a continuously flowing Krebs' solution (32–33°C, 2–3 ml/min) gassed with a 95% O2 and 5% CO2 mixture. The composition of the solution was as follows (in mm): 126 NaCl, 2.5 KCl, 1.2 NaH2PO4, 1.2 MgCl2, 2.4 CaCl2, 11 glucose, and 25 NaHCO3.
For intracellular recordings, sharp electrodes were used. They were filled with 2 m KCl (30–60 MΩ). An Axoclamp 2A amplifier (Axon Instruments, Foster City, CA) was used for recordings in either current-clamp or voltage-clamp mode. For synaptic stimulation, bipolar electrodes were used. The stimulating electrode was located in either the cortical areas close to the recording electrode or the white matter between the cortex and the striatum. As the conditioning high-frequency stimulation (HFS), we used three trains (3 sec duration, 100 Hz frequency, at 20 sec intervals).
An Axopatch 1D amplifier (Axon Instruments) was used for whole-cell patch-clamp recordings. Currents steps were generated using pClamp version 8.0 software (Axon Instruments). The striatum could be readily identified under low-power magnification, whereas individual neurons were visualized in situ using a differential interference contrast (Nomarski) optical system. This used an Olympus Optical (Tokyo, Japan) BX50WI non-inverted microscope with 40× water immersion objective combined with an infrared filter, a monochrome CCD camera (model 4912; Cohu, San Diego, CA), and a personal computer-compatible system for analysis of images and contrast enhancement (WinVision 2000; Delta Sistemi, Rome, Italy). Recordings were made with borosilicate glass pipettes (1.8 mm outer diameter; 3–5 MΩ) containing the following (mm): 125 K +-gluconate, 10 NaCl, 1.0 CaCl2, 2.0 MgCl2, 1 BAPTA, 19 HEPES, 0.3 guanosine triphosphate, and 2.0 Mgadenosine triphosphate, adjusted to pH 7.3 with KOH.
Quantitative data on post-tetanic or post-treatment modifications are expressed as percentage of the controls, the latter representing the mean of responses recorded during a stable period (15–30 min) before tetanic stimulation or drug application. Each data point in the graphs in the figures was obtained from at least five single neurons. Student's t test and χ 2 test (for paired and unpaired observations) were used to compare the means, and ANOVA was used when multiple comparisons were made against a single control group.
Drugs were applied by dissolving them to the desired final concentration in saline and by switching the perfusion from control saline to drugcontaining saline. The drugs used were as follows: APV, CNQX, SCH 23390, S-nitroso-N-acetylpenicillamine (SNAP), 7-nitroindazole monosodium salt (7-NINA) (Tocris Cookson), DA, nifedipine, quinpirole, SKF 38393 (Sigma-RBI, St. Louis, MO), H89 (Calbiochem, La Jolla, CA), and zaprinast (Rhône-Poulenc Rorer, Dagenham, UK).
Immunocytochemistry. For immunocytochemical studies, WT (n = 4) and D1-/- mice (n = 5) were deeply anesthetized with sodium pentobarbital and perfused transcardially with 4% paraformaldehyde made in 0.1 m phosphate buffer, pH 7.4. Brains were removed, postfixed in the same solution for 2 hr at 4°C, cryoprotected with 30% sucrose in 0.1 m PBS, frozen in dry ice, and sectioned (25 μm thick) in a freezing microtome. Coronal sections through the striatum were collected and stored in PBS with 0.02% sodium azide at 4°C for immunocytochemistry. All studies were approved by the appropriate animal care committee.
To localize D5 DA receptors, a polyclonal rabbit antibody against the D5 receptor subtype (1:2000 dilution) was developed and characterized at the Cajal Institute (Rivera et al., 2002a; Centonze et al., 2003). Somatostatin (SS) interneurons in the striatum correspond to neurons that express nitric oxide synthase (NOS) (Vincent et al., 1983). Therefore, to localize NOS-positive neurons, we used a monoclonal rat antiserum against somatostatin (1:100 dilution; provided by Drs. J. Rodrigo and A. C. Cuello, Cajal Institute, Consejo Superior de Investigaciones Científicas, Madrid, Spain) (Milstein and Cuello, 1983). Primary antibodies were diluted in 0.1 m PBS, with 0.2% Triton X-100 (PBS-TX), 1% bovine serum albumin (BSA), and 0.1% sodium azide.
Free-floating double-labeling immunocytochemistry was performed in sections from WT and D1-/- mice (Rivera et al., 2002a,b). Briefly, nonspecific binding sites in the sections were blocked with 5% BSA in PBS-TX for 30 min. Sections were washed and incubated with our rabbit polyclonal anti-D5 antiserum and with the monoclonal anti-SS antibody for 48 hr at 4°C. Sections were then incubated with two secondary antibodies, Alexa 598-conjugated goat anti-rabbit (red; Molecular Probes, Eugene, OR) and with goat anti-rat CY2 (green; Amersham Biosciences, Buckinghamshire, UK) diluted 1:500 in PBS-TX for 1 hr in the dark. In each experiment, the specificity of staining was monitored by incubating control sections with one primary antibody and then with both secondary antibodies to detect any cross-reaction between them. Sections were mounted in PBS/glycerol (1:1) and 2% 1,4-diazabicyclo-[2.2.2]octane (Sigma-Aldrich Química SA, Madrid, Spain), coverslipped, and observed by laser confocal microscopy (TCS-NT; Leica, Wetzlar, Germany).
Effects of D1-like receptor blockade on locomotor activity of WT and D1-/- mice
In the absence of any pharmacological challenge, D1-/- mice were significantly more active than their WT counterparts when tested for locomotor activity. According to a previous report (Xu et al., 1994), D1-/- mice interrupted photocell beams more frequently than control mice throughout the whole period of observation (Fig. 1A).
To extend our investigation to the potential role of D5 receptors on motor activity, we also analyzed the motor effects of SCH 23390, a D1/D5 receptor antagonist, in both WT and D1-/- mice. In WT mice, administration of SCH 23390 strongly suppressed motor activity, as indicated by a dose-dependent reduction of photocell beam interruptions. The percentage of reduction after 60 min of treatment was ∼30% with the lowest dose used (30 μg/kg) and ∼80% with the highest dose (300 μg/kg). Although the pharmacological effect of SCH 23390 in D1-/- mice was less pronounced, also in this genotype SCH 23390 induced a reduction of locomotor activity (35 and 45% for 100 and 300 μg/kg, respectively), suggesting a facilitatory action of D5 receptors on motor activity (Fig. 1B).
LTD and LTP in WT and D1-/- mice
Intracellular recordings from electrophysiologically identified striatal spiny neurons (Calabresi et al., 1992a, 1997, 2000; Centonze et al., 1999; Kerr and Wickens, 2001) were performed from WT and D1-/- slices during the activation of corticostriatal terminals. HFS of these terminals was able to produce long-term changes (LTD or LTP) in EPSP amplitude of control mice. As reported previously (Calabresi et al., 1997, 2000), HFS protocol was delivered in the presence of 1.2 mm external magnesium to optimize the appearance of LTD, whereas LTP induction was favored by the removal of this ion from the bathing solution. Both forms of synaptic plasticity were present in WT mice (n = 9 for LTD; n = 10 for LTP). Conversely, whereas corticostriatal LTD was still present in slices prepared from D1-/- mice (n = 11; p < 0.001), LTP was absent in these mutants (n = 13; p > 0.05) (Fig. 2). These data indicate that D1 receptors are critically important for LTP induction but are unnecessary for LTD. LTD induction, in fact, rather required D5 receptor stimulation, because this form of synaptic plasticity was blocked in D1-/- mice (n = 6), as well as in their WT counterparts (n = 6), by preincubation (7–10 min) with the D1 and D5 receptor antagonist SCH 23390 (10 μm). As reported previously (Calabresi et al., 2000), SCH 23390 was also able to fully prevent corticostriatal LTP in WT mice (n = 5) (Fig. 3A,B).
The number of striatal neurons expressing the NMDAR1 subunit of NMDA receptors is decreased in D1-/- mice (Ariano et al., 1998), and the stimulation of D1 receptors has been found to potentiate NMDA receptor responsiveness in these cells (Cepeda et al., 1998). These findings might well account for the loss of NMDA receptor-dependent LTP seen in D1-/- mice. Thus, to address this issue, we measured the amplitude, half-decay time, and duration of the membrane depolarizations induced by HFS in WT and D1-/- mice during the induction protocol of corticostriatal LTP. These parameters did not significantly differ in the two classes of animals. The amplitude was 42 ± 6mVinWT(n = 10) and 40 ± 5 mV in D1-/- mice (n = 13; p > 0.05). The half-decay time was 3.0 ± 0.4 sec in WT (n = 10) and 3.0 ± 0.6 sec in mutants (n = 13; p > 0.05). The duration was 8.1 ± 2 sec in WT (n = 10) and 8.2 ± 3 sec in mutants (n = 13; p > 0.05) (data not shown). In an additional set of experiments, we investigated the effects of the L-type calcium channel blocker nifedipine (10 μm, 10 min; n = 5) on the expression of LTP in WT mice. Although this agent has been reported to inhibit the D1 receptormediated potentiation of NMDA currents (Cepeda et al., 1998), it failed to prevent corticostriatal LTP (p < 0.01) (data not shown).
Effects of protein kinase A inhibition in corticostriatal LTP and LTD of WT and D1-/- mice
Activation of both D1 and D5 receptors results in the stimulation of protein kinase A (PKA) activity. Therefore, we tested the effects of the selective PKA inhibitor H89 on both corticostriatal LTD and LTP. In WT mice, the intracellular injection of H89 (100 μm) fully blocked LTP (n = 6; p > 0.05) but failed to prevent HFSinduced LTD in both WT mice (n = 5; p < 0.01) and D1-/- animals (n = 5; p < 0.01). In contrast, when H89 (10 μm; n = 4 for each experimental condition) was added to the perfusing solution, LTP of WT mice and LTD of both WT and D1-/- mice were prevented (p > 0.05 for both conditions) (Fig. 3C,D). These findings confirm the requirement of PKA stimulation for corticostriatal LTP and LTD in both WT and D1-/- mice and indicate that LTP induction required the stimulation of the D1–PKA pathway in the projection neuron, whereas LTD depended on the activation of D5–PKA signaling in a neuronal subtype other than the recorded spiny neuron.
Expression of D5 receptors in NOS-positive striatal neurons
The SS- and NOS-positive interneuron is a major candidate for the action of DA during the induction phase of corticostriatal LTD. Physiological studies, in fact, have suggested that intrastriatal release of nitric oxide (NO) is a critical factor for LTD induction, by promoting through a feedforward mechanism the stimulation of the cGMP–PKG pathway in spiny projection neurons (Calabresi et al., 1999). In line with this idea, DA D1-like receptor stimulation depolarizes putative NO-producing interneurons (Centonze et al., 2002a) and increases striatal levels of cGMP (Altar et al., 1990; Morris et al., 1997). The inability of D1 receptor disruption in preventing LTD, therefore, might depend on a normal sensitivity to DA of D1-/- SS-positive interneurons.
To address this issue, we performed double-labeling immunofluorescence experiments in both WT and D1-/- mice to investigate whether SS-positive interneurons (which correspond to NOS-positive neurons) (Vincent et al., 1983) of the striatum expressed D5 receptors. Although a small population of SS-positive neurons lacked D5 receptors (Fig. 4B,C, asterisks), the large majority of both WT and D1-/- SS-positive cells expressed D5 receptors (Fig. 4C,F, arrowheads). These receptors were also present in other medium- and large-sized neurons on the striatum, as shown previously (Rivera et al., 2002a).
Effects of DA receptor stimulation in low-threshold spike interneurons of D1-/- mice
SS- and NOS-positive interneurons correspond to the electrophysiologically identified low-threshold spike (LTS) cells (Kawaguchi, 1993; Kawaguchi et al., 1995). Thus, we recorded from these neurons in striatal slices of D1-/- mice. The electrophysiological features of LTS interneurons of D1-/- mice were unique among striatal cells and consisted essentially in low resting membrane potential (-58 ± 2 mV), exceptionally high input resistance (520 ± 166 MΩ, measured with patch-clamp electrodes), and in the ability to generate, in addition to fast spikes, large plateau depolarizations and low-threshold calcium spikes during depolarizing current pulses and immediately after hyperpolarization (Fig. 5A). These features did not significantly differ from those reported previously (Kawaguchi, 1993; Koos and Tepper, 1999; Centonze et al., 2002a).
In all tested D1-/- LTS cells (three of three), application of 30 μm DA (4–8 min) elicited a slow and reversible membrane depolarization, which was sufficient to initiate a train of action potentials. This excitatory effect was likely mediated by D5 receptors, because it was mimicked by the D1-like receptor agonist SKF 38393 (10 μm; n = 3; p < 0.01) but not by the D2-like receptor agonist quinpirole (10 μm; n = 3; p > 0.01) and was blocked by 10 μm SCH 23390 (Fig. 5B).
Role of the NO–cGMP pathway in corticostriatal LTD recorded from D1-/- mice
The idea that the integrity of striatal LTD in D1-/- mice depends on an intact response to DA of NO-producing interneurons predicts that, also in D1-/- mice, striatal LTD requires NO–cGMP pathway stimulation downstream D5 receptors. In D1-/- mice, therefore, we studied the effect of the NO donor SNAP (100 μm) and the selective cGMP phosphodiesterase inhibitor zaprinast (15 μm) on corticostriatal EPSP amplitude recorded in the presence of 10 μm SCH 23390. As shown in Figure 5C, incubation (7 min) with SNAP (n = 4) or zaprinast (n = 6) induced corticostriatal LTD in D1-/- slices bathed with SCH 23390 (p < 0.01 for both experimental conditions). In addition, as reported in control animals (Calabresi et al., 1999), also in D1-/- mice HFS-induced LTD was fully prevented by the NOS inhibitor 7-NINA (10 μm; n = 5; p > 0.05) (Fig. 5D).
In the present study, we provided evidence that the two members of D1-like family of DA receptors exert distinct roles in motor activity and corticostriatal synaptic plasticity. In particular, whereas genetic ablation of D1 receptors blocked LTP induction and increased locomotor activity, pharmacological blockade of D1/D5 receptors by SCH 23390 prevented LTD and inhibited motor activity in both WT and D1-/- mice.
Complex regulation of locomotor activity by D1 and D5 receptors
In D1-/- mice, locomotion was increased, whereas SCH 23390, aD1 and D5 receptor blocker, caused in the two groups of animals suppression of motor activity. SCH 23390, however, was less effective and less potent in D1-lacking animals than in control mice in inhibiting motor activity, suggesting that D1 receptors also exert a permissive action on D5 receptor function. To explain the dual role of D1 receptors in motor control, it can be speculated that D1 receptors synergize with D5 receptors when the two receptors are activated in the same cellular subtype, but they oppose to each other when activated in distinct neuronal populations. This hypothesis is supported by the following data. First, both D1 and D5 receptors are positive regulators of cAMP levels (Grandy et al., 1991; Sunahara et al., 1991; Tiberi et al., 1991; Vallone et al., 2000) and therefore cooperate in triggering common cellular events when coexpressed. In this respect, D5 receptors are also expressed in medium spiny neurons of the striatum (Rivera et al., 2002a; present study), a neuronal subtype particularly enriched in D1 receptors (Gerfen et al., 1990; Le Moine et al., 1991; Surmeier et al., 1992, 1996; Aizman et al., 2000). Second, D1 and D5 receptors are primarily expressed in different striatal cell populations (Bergson et al., 1995; Surmeier et al., 1996; Yan and Surmeier, 1997; Rivera et al., 2002a), suggesting that they can also exert distinct physiological actions. Noticeably, NOS-positive neurons, which express D5 receptors and are stimulated by D1-like receptor agonists, do not have D1 receptors (Le Moine et al., 1991; Kawaguchi et al., 1995; Rivera et al., 2002a) and cause longterm inhibition of the excitability of striatal spiny neurons, thereby contrasting the direct LTP-favoring effects of D1 receptors on these cells. In this line, DAergic activation of NOS-positive cells causes a long-term inhibition of medium spiny neurons detected by a reduction of c-Fos expression in striatal matrix neurons (Moratalla et al., 1996a,b).
Other interpretations of our results are also possible. Accordingly, the postulated inhibitory role of D1 receptors only emerges from the increased motor activity seen in D1-/- mice, although it could in principle result from a variety of developmental or compensatory changes. In addition, the evidence that the D1/D5 receptor antagonist SCH 23390 inhibited locomotor activity in both WT and D1-/- and that this action was more pronounced in WT might also suggest that D1 receptors exert only stimulatory actions on locomotor activity by synergizing with D5 receptors. The use of the recently generated mice lacking D5 receptors (Holmes et al., 2001; Hollon et al., 2002) could help to clarify this important issue.
Complex regulatory role of DA receptors in corticostriatal synaptic plasticity
Repetitive stimulation of corticostriatal pathway can induce either LTD or LTP both in vivo (Charpier and Deniau, 1997; Reynolds and Wickens, 2000) and in vitro (Calabresi et al., 1992a,b; Dos Santos Villar and Walsh, 1999; Partridge et al., 2000). Persistent changes in synaptic strength in the striatum are considered as neural correlates of specific motor abilities, although more recently have been involved in several other aspects of brain activity, such as reward-related learning (Berke and Hyman, 2000; Reynolds et al., 2001), maturation of neural circuitry during development (Choi and Lovinger, 1997), and drug addiction (Berke and Hyman, 2000; Hyman and Malenka, 2001; Nestler, 2001).
A solid achievement in the field of basal ganglia physiology is the requirement of DA for corticostriatal synaptic plasticity. The unavailability of pharmacological agents able to target selectively specific members of the two subfamilies of DA receptors, however, makes essential the use of mutant mice to identify the receptor subtypes involved in these physiological activities. On the basis of this consideration, mice lacking DA D2 receptors have been used previously for electrophysiological experiments in vitro. These experiments allowed to identify the D2 receptor as the specific member of the D2-like subfamily involved in the facilitatory effect on LTD and in the restraining action on LTP (Calabresi et al., 1997). Together, these results with the present data and with the observations that both D1- and D2-like antagonists block locomotor activity (Vallone et al., 2000) and striatal LTD (Calabresi et al., 1992a,c), it is possible to conclude that D5 and D2 receptors cooperate functionally to facilitate motor activity and striatal LTD, whereas D1 and D2 receptors are the receptor subtypes mainly involved in striatal LTP and, possibly, motor inhibition. It is also possible, however, that the role of D5 receptors in LTD only emerges as a compensatory mechanism after D1 receptor ablation and that it is negligible in control conditions. Again, the use of D5 receptor-lacking mice will prove useful to clarify this issue.
Current models of the basal ganglia organization propose that the striatum is an important component of motor, cognitive, and limbic circuits. This nucleus takes part in several brain activities by processing the flow of information arising from different neocortical areas and projecting to the thalamus (Bergman et al., 1998). DA plays a crucial role in these processes by affecting the activity of striatal cells through multiple mechanisms. Interestingly, a complex interplay between different subtypes of DA receptors is required for both striatal motor control and synaptic plasticity, possibly involving different subpopulations of striatal neurons.
Investigating the receptor and cellular mechanisms involved in striatal motor control and synaptic plasticity is an essential requirement to understand the neural bases of critical brain functions and of dramatic pathological conditions, such as Parkinson's disease, schizophrenia, and drug addiction.
Correspondence should be addressed to either of the following: Paolo Calabresi, Clinica Neurologica, Dipartimento di Neuroscienze, Università di Tor Vergata, Via Montpellier 1, 00133 Rome, Italy, E-mail:; or Rosario Moratalla, Instituto Cajal, Consejo Superior de Investigaciones Científicas, Avda Dr. Arce 37, 28002 Madrid, Spain, E-mail:
Copyright © 2003 Society for Neuroscience 0270-6474/03/238506-07$15.00/0
This work has been supported by Ministerio de Ciencia y Tecnologìa Grant SAF200-0122, Plan National Sobre Drogas, and Fundación La Caixa (Spain) (R.M.) and Consiglio Nazionale delle Ricerche and Ministero della Salute, Progetto Finalizzato Schizofrenia (Italy) (P.C.). We thank Sonia Villa for her help with the behavioral experiments.
↵* R.M. and P.C. contributed equally to this work.