Additional contributions could also arise from the enhanced recovery of voltage-gated Na+ channels from inactivation (Aman and Raman, 2007). Rebound firing in direct response to synaptic inhibition has been proposed (and is widely accepted as an obvious mechanism) but it has rarely been demonstrated (Nambu and Llinas, 1994) particularly in response to physiological sensory stimulation in the mammalian CNS. Recent studies employing synaptic stimulation were unable to demonstrate physiological rebound firing in the deep cerebellar nuclei (Alviña et al., 2008). In songbirds rebound firing has been linked

with vocal learning, where thalamic neurons translate IPSPs into an excitatory output (Bottjer, 2005 and Person and Perkel, 2005) and modeling studies clearly show VX-770 cost the potential for IH to generate rebound firing in the mammalian brain, but the key physiological question is: how can a physiological input sufficiently activate IH to generate this firing? Here we show that the SPN uses powerful chloride extrusion selleck compound to extend the physiological voltage range negative to EK. This enhances the chloride driving force of IPSPs, which can then provide sufficient hyperpolarization to activate the IH conductance. IH has a general role in modulating

input resistance and hence the membrane time constant; this is especially important in the auditory system, which depends on speed and temporal precision (Bal and Oertel, 2000 and Oertel et al., 2008). Although sound localization through mechanisms accurately discriminate submillisecond

time intervals (McAlpine et al., 2001), the MNTB-SPN circuit forms an early computation adapted to encode millisecond to second time intervals. The idea that IH could be involved in this computation was first proposed from the modeling studies of Hooper et al. (Hooper et al., 2002), who suggested different cell categories (low-pass, band-pass, or high-pass) to encode sound of different durations, but all limited to sounds lasting longer than 50 ms. For instance, induction of offset responses in the IC by 200 ms hyperpolarizing current injections was mediated by IH (Koch and Grothe, 2003), while 50 ms sound pulses failed to do so in the same nucleus (Xie et al., 2007). However, encoding derives not only from stimulus duration but also from “intensity,” since loud sounds with higher input firing rates will generate greater summation of IPSPs and activate more IH current. Therefore, a short-duration sound could elicit an offset response if delivered at a higher intensity, and provided the activation kinetics of IH were fast enough. Coincidence-based modeling of IPSPs and EPSPs (Aubie et al.

There was much discussion within the professionals group about CM being considered as a politically driven initiative,

which was extended to more general feelings of antipathy towards treatment guidelines. There is a substantial literature on the length of time that it takes to get new research adopted into practice (Benishek et al., 2010 and McGovern et al., 2004). Although not all the groups were aware of the literature of the effectiveness of CM, there was a general assumption that a literature existed as the basis of a national guideline. However, as with other studies, practitioners were PF-01367338 datasheet quick to cite that the research evidence did not reflect the complexity of the service users or clinical situation of routine practice and this affects the perception of its usefulness for clinical decision making (Miller, 1987, Greenhalgh et al., 2004, Kirby et al., 2006 and Pilling et al., 2007). The study is limited by the relatively small number of participants who took part in the focus groups. Issues of generalisability have a different focus within qualitative work, in that a study of this kind seeks to raise awareness of the concepts and define the phenomena to be further refined and tested for prevalence using other methods (Craig et al., 2008). The smallest focus group only included two female service users (both working as prostitutes) but the relative privacy of

this group allowed Regorafenib manufacturer for an in-depth exploration of the issues that they may have been less happy to engage with in a larger group. One of the strengths of this study is that the use of qualitative methods allows for a more in-depth and contextualised exploration of the factors which may influence the implementation and effectiveness of a complex intervention such as CM. Previous studies have shown that there are differences in the attitudes of staff members to CM (Benishek et al., 2010, Kirby et al., 2006, McGovern et al., 2004 and Petry, 2006) and this study highlights the complex interaction of professional attributes much and

personal beliefs that may underlie these attitudes. That many of the concerns about CM appear to be similar in this smaller number of UK practitioners to the larger US surveys (Benishek et al., 2010 and McGovern et al., 2004) suggests the validity and generalisability of these results, and some common cross cultural themes that require more robust process evaluation in future RCTs. A final strength of this study is the inclusion of service users within the analysis, and the different emphasis that they bring to treatment decision making. There is a growing literature demonstrating the importance of including process evaluation as an essential part of clinical trials of complex interventions (Audrey et al., 2006, Craig et al., 2008, Hawe et al., 2004 and Lewin et al., 2009).

, 2000 and Lewis et al., 2004) and Dlx I12b-Cre. The Dlx I12b-cre allele is expressed in SVZ and MZ, but not the VZ, of the basal ganglia, beginning around E10.5 ( Potter et al., BMS 754807 2009). We refer to this mutant as Dlx1/2-cre;ShhF/−. The Dlx1/2-cre;ShhF/− mutant lacked expression of Shh exon 2 RNA and SHH protein (in the MGE MZ [arrows, Figures 4B, 4B′, 4C, and 4C′], but not in the VZ [arrowheads, Figures 4B, 4B′, 4C, and 4C′, and not shown]). On the other hand, Shh transcripts continued to be expressed in the Dlx1/2-cre;ShhF/− MGE MZ ( Figures 4D–4F′), showing that Shh expression in these cells does not require continued production of Shh protein, and that this cell type was present. At E11.5,

the Dlx1/2-cre;ShhF/− exhibited reduced MGE SHH signaling based on ∼2-fold decreased Gli1, Ptc1, and Nkx6-2 expression in the VZ and SVZ of the overlying MGE MZ compared to control brains (Cre−;ShhF/− or Cre−;ShhF/+) ( Figures 4 and S4; Table S1), thus showing that Shh expression in postmitotic neurons regulates properties of MGE progenitor cells. These effects were prominent in the dorsal MGE, where Gli1 and Nkx6-2 expression were strongly reduced (arrows, Figures 4K, 4J′, 4N, and 4M′). Ptc1, Gli1, and Nkx6-2 expression in the ventral-most MGE and preoptic area appeared normal, presumably because Shh expression in the VZ was not affected. Surprisingly, NKX2-1

expression was largely unchanged, except for the loss of expression in the VZ of the dorsal-most MGE (arrows, Figures

4Q and 4Q′). Furthermore, the mutant’s MGE did not show an obvious morphological change. The Dlx1/2-cre;ShhF/− mutation preferentially altered differentiation DAPT solubility dmso in the rostrodorsal MGE at E14.0 ( Figure 5). Ptc1 expression in the VZ of this region was greatly reduced, whereas its expression in the preoptic area remained strong ( Figure S5). Likewise, expression of Gli1, Lhx6, Lhx8, Nkx2-1, and Nkx6-2 in the VZ, SVZ, and MZ were selectively reduced Isotretinoin in the rostrodorsal MGE (arrows, Figures 5 and S5). Consistent with this, properties of the mutant’s globus pallidus appeared unchanged (Er81, Lhx6, Lhx8, Nkx2-1, and Zic1; Figure S5), as the globus pallidus is largely derived from the ventral MGE and dorsal preoptic area ( Flandin et al., 2010 and Nóbrega-Pereira et al., 2010). The rostrodorsal MGE of the Dlx1/2-cre;ShhF/− mutant had strong phenotypes at E18.5 ( Figures 6 and S6). This area includes the region of the anterior extension of the bed nucleus of stria terminalis (medial division; STMA) and the core of the nucleus accumbens (AcbC). These regions lacked detectable expression of Lhx6, Lhx8, and Nkx2-1 in the progenitor zone and showed reduced expression in the mantle zone (arrows, Figure 6); Lhx6 expression in tangentially migrating cells coursing through the LGE SVZ was also reduced (arrowheads, Figures 6C and 6C′). On the other hand, expression of Islet1 in the AcbC appeared normal ( Figure S6).

Protein translation plays a major role in mGluR5 signaling (Lüscher and Huber, 2010). The phosphorylation of eEF2 is increased by Aβo-PrPC as much as by mGluR5 agonist. Therefore dysregulation of translation may contribute to synaptic dysfunction in AD. Arc is one protein target of mGluR5 signaling that is upregulated by Aβo acutely. Calcium find more and Fyn are independent mediators, which appear to cooperate in eEF2 phosphorylation. We show that mGluR5 antagonists prevent Aβo-induced spine loss from hippocampal neurons in vitro and in vivo. Critically, MTEP reverses memory deficits in transgenic AD models. Multiple signaling pathways from

Aβo-PrPC-mGluR5 complexes are likely to participate. For spine loss in vitro, Fyn is required (Um et al., 2012), but other mGluR5 signaling components may contribute. Protein translation, calcium release, and Fyn kinase are each known to participate in plasticity, learning, and memory. The mGluR5 pathway may also feedback on APP/Aβ metabolism to exacerbate AD. Specifically, mGluR5 agonism elevates Arc, which enhances Aβ production by participating in APP and PS1 colocalization within endocytic LGK-974 mw vesicles (Wu et al., 2011). Shared pathways between AD and Fragile X have been reported (Sokol et al., 2011). The FMRP protein normally represses APP translation. Transgenic mice with both APP transgenes and loss of FMRP have enhanced phenotypes, including

audiogenic seizures, which are treatable with MPEP. Of mGluR receptors, only mGluR1 and mGluR5 interact with Fyn and PrPC. Only mGluR5 mediates Aβo-induced stimulation of Fyn and calcium signaling in oocytes. Grm5 gene deletion

and mGluR5-specific compounds reverse Aβo phenotypes, including Fyn activation, neuronal calcium mobilization, eEF2 phosphorylation, spine loss, LDH release, and memory deficits. The mGluR1-specific antagonist, MPMQ, does not block. Thus, mGluR5 appears to be specifically involved in Aβo-PrPC action. PrPC, mGluR5, and Fyn have all been localized to the PSD by subcellular fractionation. For PrPC and Fyn, high-resolution in situ protein localization in brain has not been reported. Resminostat For mGluR5, imaging confirms a postsynaptic localization and indicates that mGluR5 is dynamically located at the PSD periphery (Lujan et al., 1996). Dynamic regulation of mGluR5 localization by Aβo has been observed (Renner et al., 2010). Although ionotropic receptors function rapidly, metabotropic receptors are slow and show prominent desensitization. Aβo levels are highly unlikely to fluctuate on the time scale of synaptic transmission, so Aβo-PrPC complexes may engage mGluR5 and elicit a degree of desensitization that prevents responsiveness to cyclic changes in Glu. Thus, mGluR5 may be dysregulated by acute activation and chronic desensitization. Activation of mGluR5 by Aβo-PrPC complexes expands the repertoire of metabotropic glutamate receptors.

, 2006 and Takashima et al., 2009; Takehara-Nishiuchi and McNaughton, 2008). The proposition that VMPFC is recruited after initial memory integration by hippocampus HDAC inhibitor is also supported by the fact that hippocampal, but not VMPFC, encoding activation predicted inference success in single-trial associative

learning (Zeithamova and Preston, 2010). Alternatively, VMPFC increases in the present study may reflect organization or resolution of overlapping memory representations (Hasselmo and Eichenbaum, 2005; Ross et al., 2011) that leads to their integration in the current paradigm. Increased hippocampal-VMPFC functional coupling across repetitions of overlapping events Vemurafenib chemical structure was accompanied by corresponding increases in hippocampal functional connectivity with precuneus, superior parietal cortex, and frontal pole. These regions—along with the hippocampus and VMPFC—are considered part of the default network (Raichle et al., 2001) that is also engaged during simulation of future events (Addis et al., 2009; Andrews-Hanna et al., 2010) and successful episodic remembering (Buckner et al., 2005; Greicius et al., 2004), in particular during

recollection of specific event details (Vincent et al., 2006). Based on this evidence, it has been proposed

that the default network supports the formation of mental models of significant events, particularly when judgments about those events depend on inferred content (Buckner et al., 2008). The default network would support the these reactivation of prior events that could then be recombined and recoded into prospectively useful models of experience (Buckner, 2010). The present findings provide support for this hypothesis, demonstrating increased coupling between hippocampus and other components of the default network during retrieval-mediated formation of relational memory networks. Several leading theories hypothesize that the fundamental flexibility of episodic memory results from our ability to form networks of related memories that link discrete events (Buckner, 2010; Cohen and Eichenbaum, 1993; O’Keefe and Nadel, 1978; Tolman, 1948). Despite the theoretical importance of this question, much empirical memory research has focused solely on encoding processes that mediate memories for individual events. Here, we demonstrate that memories for distinct experiences are integrated through a retrieval-mediated encoding mechanism, with prior related memories being reactivated and bound to the current experience during encoding.

, 2010). To test whether PrP and α2δ-1 interacted physically, we immunoprecipitated PrP from cerebellar extracts of Tg(WT) and Tg(PG14) mice, and immunoblotted the precipitated fractions with an antibody raised against the α2 polypeptide of α2δ-1. As shown in Figure 5A, an immunoreactive band of ∼145 kDa was detected in immunoprecipitates of Tg(WT) and Tg(PG14)

but not in Prnp0/0 mice, or when the immunoprecipitation was done in the absence of the anti-PrP antibody. After deglycosylation with PNGaseF, this band shifted to an apparent molecular check details weight of 107 kDa, as expected for the α2 polypeptide ( Figure S5A) ( Davies et al., 2006). The interaction was confirmed in the reverse experiment in which α2δ-1 was immunoprecipitated RG7204 purchase from cerebellar extracts and PrP detected by immunoblot (Figure 5B), and was also seen in primary cultured CGNs (Figure S5B) and transiently transfected HeLa cells (Figure S5C). HC-deleted PrP molecules coimmunoprecipitated with

α2δ-1 (Figure S5C), indicating that PrP region 114–121 was not essential for the interaction. Next, we tested whether the distribution of α2δ-1 was altered in cells expressing PG14 PrP. HeLa cells were cotransfected with plasmids encoding the CaVα1A, CaVβ4, and α2δ-1 subunits, and either wild-type or PG14 PrP-EGFP fusion proteins, and analyzed by confocal microscopy after immunofluorescent staining of α2δ-1. Consistent with previous localization of nonfluorescent and EGFP-fused PrPs (Biasini et al., 2010, Fioriti et al., Non-specific serine/threonine protein kinase 2005 and Ivanova et al., 2001), the majority of wild-type PrP localized on the cell surface (Figures 6A and 6J), whereas PG14 PrP was mostly found in intracellular

compartments (Figures 6D and 6J). In cells expressing wild-type PrP, α2δ-1 was efficiently expressed on the plasma membrane where it colocalized with PrP (Figures 6B, 6C, and 6K). In contrast, α2δ-1 was weakly expressed on the surface of PG14 PrP-expressing cells, and was mostly found in perinuclear patches where it colocalized with PrP (Figures 6E, 6F, and 6K), and with ER and Golgi markers (data not shown). This was seen in cells with high or low expression levels, ruling out that the abnormal localization of α2δ-1 was due to overexpression. The CaVα1A pore-forming subunit also accumulated intracellularly in PG14 PrP-expressing cells (Figures S6A–S6H), whereas there was no effect on the localization of 5′ nucleotidase (5′NT), a raft-resident GPI-anchored protein that does not belong to the VGCC complex (Davies et al., 2010) (Figures S6I–S6P). In cells expressing PG14/ΔHC PrP, α2δ-1 was more efficiently delivered to the cell surface, indicating that intracellular retention of mutant PrP played a role in the trafficking defect (Figures 6H, 6I, and 6K).

9 ± 0.7 s in syp−/−; ΔC-syp, τ = 20.4 ± 0.9 s in syp−/−; wt-syp) ( Figure 3A). We then examined vesicle retrieval during stimulation using the same protocol as in Figure 2A ( Figure 3B). As compared to

wt-syp, the truncation mutant syp failed to rescue defective endocytosis during neuronal activity in terms of rate (0.0095 AU s−1 in syp−/−; wt-syp, 0.0045 AU s−1 in syp−/−; ΔC-syp) ( Figures 3C and 3D) and the relative magnitude of vesicle Alectinib solubility dmso retrieval (0.28 ± 0.03 in syp−/−; wt-syp, 0.14 ± 0.03 in syp−/−; ΔC-syp) ( Figures 3B and 3E). These results suggest that the C-terminal domain of syp is selectively required for the endocytosis that occurs during, but not after, cessation of sustained synaptic transmission. A previous study reported that the C-terminal tail is essential for internalization of syp in fibroblasts (Linstedt and Kelly, 1991). We tested this notion using full-length pHluorin-tagged synaptophysin (fl sypHy) and the mutant sypHy (ΔC-sypHy) that lacks the same C-terminal segment (amino acids 244–307). ΔC sypHy fluorescence, at the end of the 10 Hz stimulation protocol

(30 s), showed a punctate distribution that was indistinguishable from full-length sypHy, reflecting efficient targeting to SVs (Figure S2A). The poststimulus endocytic time-constant of ΔC sypHy (τ = selleck chemical 18.8 ± 0.8 s) was not significantly different from full-length sypHy (τ = 18.0 ± 0.8 s) (Figure S2B), indicating that the C-terminal tail of syp is not required for efficient internalization of syp after neuronal activity. Next, we tested whether trafficking of syp, during neuronal activity, was altered in ΔC sypHy using the same protocol as in Figure 2A. Interestingly, retrieval of ΔC sypHy during neuronal activity was significantly reduced (0.31 ± 0.02 for fl sypHy, 0.18 ± 0.04 for ΔC sypHy) and also became slower as compared to full-length sypHy (0.015 AU s−1

for fl sypHy, 0.010 AU s−1 for ΔC sypHy) (Figures S2C and S2D). Thus, these results further demonstrate that different motifs within syp are involved in controlling the endocytosis of SV that occurs during, versus after, sustained synaptic transmission, potentially by recruiting distinct ensembles of proteins for recycling. We investigated many the physiological significance of the endocytic defects in syp−/− neurons by performing whole-cell voltage-clamp recordings in dissociated cortical neurons. We locally stimulated neurons by delivering electrical pulses to the cell body using a stimulating electrode and recorded evoked inhibitory postsynaptic currents (IPSCs) from the cell body of postsynaptic partners. This method has been used to examine the dynamics of SV pools in numerous studies ( Chung et al., 2010 and Ferguson et al., 2007). We measured the amplitude and the kinetics of single IPSCs between wild-type and syp−/− neurons, and found that these parameters were not altered ( Figures S3A and S3B).

Strikingly, KPT-330 ic50 recordings from single dopaminergic neurons in the ventral tegmental area (VTA) and substantia nigra pars compacta (SNc) report activity that resembles this precise error function (Schultz et al., 1997 and Waelti et al., 2001). Dopamine neurons signal unpredicted rewards but are silent when rewards are fully predicted, instead firing at the occurrence of the earliest predictive stimulus. When an expected reward is omitted, dopamine neurons depress their activity at the precise time that this reward should have occurred. Hence, when stimulus-outcome

associations are precise in time, dopaminergic activity, like the TD error function, is precise in time (Hollerman and Schultz, 1998). By comparison, little is known about dopaminergic activity when the time between predictive event and resulting reward is imprecise. When the occurrence of reward is fully predicted, dopamine neurons show differential firing for equal rewards occurring at different times (Hollerman and Schultz, 1998 and Fiorillo et al., 2008). A similar dependence of an RPE on the precise time of reward delivery in the case of unpredicted or partially predicted rewards would have implications for the role of dopamine in learning. More specifically, such a signal

would be most relevant in situations where the goal is to learn not only how much, but also precisely when, a reward will ensue. A temporal dependence for a dopaminergic RPE signal would also have implications for understanding striatal activity as measured by BOLD fMRI, where numerous of studies report a correlation between the BOLD signal and RPE in learning studies (O’Doherty et al., 2003, Tobler et al., 2006, selleck inhibitor Pessiglione et al., 2006, Schönberg et al., 2007 and Valentin and O’Doherty, 2009). Although it is possible to detect RPE correlates in the VTA (D’Ardenne et al., 2008), technical limitations imaging this region have meant that it is consistently easier to test for such signals in the striatum.

Indeed, a large VTA/SNc projection to the striatum has fostered an implicit assumption that activity here reflects a dopaminergic input (O’Doherty et al., 2004 and Campbell-Meiklejohn et al., 2010; and many similar examples). In fMRI studies, it is often advantageous to introduce significant temporal jitter between events. Whereas some researchers have chosen to eschew this advantage in favor of maintaining temporal precision (Schönberg et al., 2010, O’Doherty et al., 2003, Pessiglione et al., 2006, Gershman et al., 2009 and Krugel et al., 2009), others have chosen to maximize BOLD signal sensitivity by introducing significant randomness (up to 10 s) in the interval between conditioned stimulus and outcome (Behrens et al., 2007, Behrens et al., 2008, Hare et al., 2008, Cohen et al., 2010 and Daniel and Pollmann, 2010). This temporal jitter has in all cases been ignored in the computation of the prediction error, subsequently found to correlate with striatal BOLD signal.

This work was supported by the National Institute of Aging (NIA grants AG19724 and AG1657303 to B.L.M. and W.W.S.), the Larry L. Hillblom Foundation (W.W.S. and J.H.K.), and the John Douglas French Alzheimer

Foundation (W.W.S.), and the Consortium for Frontotemporal Dementia Research. We thank our research participants and their families for contributing to neurodegeneration research. “
“When we visually track a moving object with eye movements, the world around us appears still despite the self-induced retinal motion, demonstrating the remarkable capability of the visual system to integrate retinal motion signals with nonretinal signals during eye movements (Gibson, 1954, Ilg et al., 2004 and Royden et al., 1992). A failure of this integration leads to the false perception of environmental motion during eye movements as observed in a patient with bilateral parieto-occipital lesions (Haarmeier Decitabine et al., 1997). Single-unit studies in

the macaque have shown the presence of so-called “real-motion” neurons in several cortical regions that receive efference signals of eye or head movements, such as V3A, MST, VIP, V6, and the visual posterior sylvian (VPS) selleck inhibitor area (Dicke et al., 2008, Erickson and Thier, 1991, Galletti et al., 1990, Ilg et al., 2004 and Zhang et al., 2004). These neurons respond to moving stimuli during fixation, but reduce or abolish responses when retinal motion is induced by active pursuit over a static target, and are thought to mediate perceptual stability during visual pursuit. In the human brain, comparably little is known about this type of “objective” or head-centered motion response. Among motion-responsive regions V5/MT, MST, V3A, medial parietal and cingulate regions (Morrone et al., 2000, Orban et al., 2003, Tootell et al., 1997 and Wall

and Smith, 2008), MST, CSv, and putative VIP homologs have been shown to prefer complex motion types compatible with egomotion such as 3D forward-flow or full-field planar motion (Bartels et al., 2008b, Fischer et al., 2011, Morrone et al., 2000, Peuskens et al., 2001 and Wall and Smith, 2008), and to integrate visual motion signals across nonvisual modalities (Sereno and aminophylline Huang, 2006 and Smith et al., 2011). In particular, V5/MT, MST, V3A, and V6 have been similarly implicated in the integration of eye movement signals with heading-related forward flow (Arnoldussen et al., 2011 and Goossens et al., 2006), as well as in spatiotopic responses at fixed eye positions (Crespi et al., 2011 and d’Avossa et al., 2007). However, prior human studies have not examined the neural substrates involved in integrating pursuit eye movements with planar motion, which involves neural substrates that are distinct from those involved in processing heading-related expansion flow (Duffy and Wurtz, 1995, Gu et al., 2008, Morrone et al., 2000, Royden and Vaina, 2004 and Zhang et al., 2004).

Importantly, under this interpretation, new state formation is intact; however, retrieval of appropriate states is disrupted or at least less selective. A second possible explanation (option B in Figure 1, bottom) is that the rats with disrupted cholinergic function might have been able to form a new state in extinction but not in the other challenges. Why would this happen? To answer this, it is useful to ask how the brain knows that a new state should be formed in the first place. One impetus for state

creation is significant differences between the current situation and past experience (Gershman et al., 2010). According to this idea, prediction errors—differences between what is expected (driving is on the right of the road, mass transportation is called “subway,” etc.) and what is currently experienced selleck inhibitor (cars are on the left, the underground train is “the city circle”)—drive state formation. Importantly,

Afatinib mw these prediction errors include both errors in predicted value (the city circle is not cheap), and errors in predicted identity (would you expect “the city circle” to indicate an underground train system?). The former are typically termed reward prediction errors (though we use “value,” as changes in rewarding events can also induce identity prediction errors), and Bradfield et al. (2013) refer to the latter as “state prediction errors,” though we prefer “identity,” as any sort of error could lead to recognition of state change. Bradfield et al.’s first two Liothyronine Sodium manipulations—contingency degradation and reversal learning—involved only identity prediction errors,

since the underlying value of the reward associated with lever pressing did not change. However, the last manipulation introduced value prediction errors since the reward was entirely omitted. If cholinergic transmission in the striatum is important for detecting, representing, or learning from identity prediction errors, one would expect to see no new state formation in the first two manipulations due to the cholinergic manipulation, but intact state formation during extinction learning. Thus, like a retrieval deficit, a selective effect on the formation of new states following identity prediction errors would also produce the observed pattern of results (Figure 1, bottom). Though relatively little is known about the function of cholinergic striatal interneurons, what we know so far relates nicely to these two interpretations. For example, one can easily imagine a key role for striatal acetylcholine (Ach) in retrieval: cholinergic interneurons are inhibitory, tonically active, and innervate (and receive input from) a large number of medium spiny neurons (Zhou et al., 2002). This places this local modulatory system in a prime position to provide network-wide inhibition, promoting retrieval of only the relevant state at each point in time (Apicella, 2007). By reducing cholinergic tone, Bradfield et al.