Inactivation of the interpositus nucleus during unpaired extinction does not prevent extinction of conditioned eyeblink responses or conditioning-specific reflex modification

Abstract

For almost 75 years, classical eyeblink conditioning has been an invaluable tool for assessing associative learning processes across many species, thanks to its high translatability and well-defined neural circuitry. Our laboratory has adapted the paradigm to extensively detail associative changes in the rabbit reflexive eyeblink response (UR), characterized by post-conditioning increases in the frequency, size, and latency of the UR when the periorbital shock unconditioned stimulus (US) is presented alone, termed conditioning-specific reflex modification (CRM). Because the shape and timing of CRM closely resembles the conditioned eyeblink response (CR) to the tone conditioned stimulus (CS), we previously tested whether CRs and CRM share a common neural substrate, the interpositus nucleus of the cerebellum (IP), and found that IP inactivation during conditioning blocked the development of both CRs and the timing aspect of CRM. The goal of the current study was to examine if extinction of CRs and CRM timing, accomplished simultaneously with unpaired CS/US extinction, also involves the IP. Results showed that muscimol inactivation of the IP during extinction blocked CR expression but not extinction of CRs or CRM timing, contrasting with the literature showing IP inactivation prevents CR extinction during CS-alone presentations. The continued presence of the US throughout the unpaired extinction procedure may have been sufficient to overcome IP blockade, promoting plasticity in the cerebellar cortex and/or extra-cerebellar components of the eyeblink conditioning pathway that can modulate extinction of CRs and CRM timing. Results therefore add support to the distributed plasticity view of cerebellar learning.

1. Introduction

For nearly three quarters of a century, classical conditioning of the eyeblink response has been used as a tool to investigate associative learning mechanisms across many species, including humans (for review, see Woodruff-Pak & Steinmetz, 2000), cats (e.g. Gruart, Blazquez, & Delgado-Garcia, 1994; Hesslow, 1994; Woody & Brozek, 1969), rabbits (e.g. Berger, Alger, & Thompson, 1976; Gormezano, Schneiderman, Deaux, & Fuentes, 1962), and rodents (e.g. Chen, Bao, Lockard, Kim, & Thompson, 1996; Hughes & Schlosberg, 1938; Jirenhed, Bengtsson, & Hesslow, 2007; Stanton, Freeman, & Skelton, 1992). Due to the elegant simplicity yet endless variation that can be applied to the paradigm and the extensive literature detailing the underlying neural circuitry (Ammann, Marquez-Ruiz, Gomez-Climent, Delgado-Garcia, & Gruart, 2016; Cheng, Disterhoft, Power, Ellis, & Desmond, 2008; Christian & Thompson, 2003; Freeman, 2015; Hesslow & Yeo, 2002; Mauk, Li, Khilkevich, & Halverson, 2014; Yang, Lei, Feng, & Sui, 2015), it is arguably one of the most translatable learning tasks, easily adapted from animal models to the clinic and vice versa (Greer & Thompson, 2017). Although most studies primarily focus on the development of a conditioned eyeblink response (CR) to a conditioned stimulus (CS) paired with a blink-eliciting unconditioned stimulus (US), another area of interest concerns the development of associative changes in the unconditioned response (UR) (Canli, Detmer, & Donegan, 1992; Schreurs, Oh, Hirashima, & Alkon, 1995; Weisz & LoTurco, 1988). Eyeblink conditioning therefore can be used to not only examine the processes behind acquisition, retention, and extinction of CRs, but also learning-related changes in reflexive responding.

Earlier eyeblink conditioning studies, primarily in rabbits, documented associative changes in the UR that occurred in the presence of the CS, such as reflex facilitation and conditioned diminution. Reflex facilitation is an increase in UR amplitude that develops rapidly at the start of conditioning and is thought to be one of the earliest behavioral manifestations of associative learning (Harvey, Gormezano, & Cool-Hauser, 1985; Lam, Wong, Canli, & Brown, 1996; Weisz & McInerney, 1990; Young, Cegavske, & Thompson, 1976). In contrast, conditioned diminution, a decrease in UR amplitude, can also occur when a trained CS precedes a US, a phenomenon that is not simply the result of response habituation (Canli et al., 1992; Donegan, 1981; Kimble & Ost, 1961). Our laboratory has documented changes in the UR that occur in the absence of the CS when the US is tested by itself, termed conditioning-specific reflex modification (CRM) (for review, see Burhans, Smith-Bell, & Schreurs, 2008; Schreurs, 2003; Schreurs & Burhans, 2015). CRM is characterized by increases in the frequency, latency, amplitude, and area of URs following delay eyeblink conditioning in rabbits, particularly to USs that are four to eight folds smaller than the US utilized during CS-US pairings (Schreurs et al., 1995). CRM is deemed a conditioning-specific because the same changes in the UR are not found in subjects receiving explicitly unpaired CS and US presentations (Buck, Seager, & Schreurs, 2001; Schreurs et al., 1995; Schreurs, Shi, Pineda, & Buck, 2000; Schreurs, Smith-Bell, & Burhans, 2018). CRM-like changes have also been reported by others in rabbits (Gruart & Yeo, 1995; Wikgren & Korhonen, 2001; Wikgren, Ruusuvirta, & Korhonen, 2002) and rodents (Servatius, Brennan, Beck, Beldowicz, & Coyle-DiNorica, 2001).

Initial studies of CRM identified the training parameters that promote its development and extinction. In brief, these studies revealed that CRM levels increased as the number of training sessions or intensity/aversiveness of the conditioning US increased, parameters which also can affect CR levels (Buck et al., 2001; Schreurs et al., 1995; Seager, Smith-Bell, & Schreurs, 2003). UR response topographies were also found to resemble the CR response topography during CS-US pairings, suggesting that CRM may be a CR that has generalized from the CS to the US (Schreurs, 2003). Subsequent studies then focused on finding treatments, both behavioral and neurochemical, that may induce extinction of CRM along with CRs (Burhans, Smith-Bell, & Schreurs, 2013, 2015, 2017, 2018; Schreurs et al., 2000). A dichotomy between CRs and CRM became clear as it was found that CRs and CRM were best extinguished by CS-alone and US-alone presentations, respectively, with simultaneous extinction of CRs and URs only occurring when these two treatments were combined as unpaired extinction (Schreurs et al., 2000). Importantly, extinction of CRM occurred even if the US-alone presentations during unpaired extinction sessions were significantly reduced in intensity from the US used during conditioning (Schreurs, Smith-Bell, & Burhans, 2011). Later studies showed that CRs and CRM differentially responded to systemic serotonergic, glutamatergic, and noradrenergic manipulations, further highlighting a differentiation between the two (Burhans et al., 2013, 2017, 2018). Overall, these studies suggested that CRs and CRM may share some overlapping neural substrates but also some distinct neural mechanisms.

Delay eyeblink conditioning is a cerebellar-dependent learning task, with strong evidence for two major sites of plasticity, the cerebellar cortex and the deep cerebellar nuclei, primarily the interpositus nucleus (IP) (for review, see Christian & Thompson, 2003; Freeman, 2015). Anatomical and electrophysiological data have pinpointed the IP as one of the major sites of convergence of CS and US input pathways with CR motor output pathways (Gonzalez-Joekes & Schreurs, 2012; Gould, Sears, & Steinmetz, 1993; Ostrowska, Zguczynski, & Zimny, 1992; Steinmetz & Sengelaub, 1992; Thompson, 2013), and many lesion and inactivation studies have demonstrated that the IP is necessary for the acquisition and expression of CRs (e.g. Bracha, Webster, Winters, Irwin, & Bloedel, 1994; Burhans & Schreurs, 2018; Clark, McCormick, Lavond, & Thompson, 1984; Garcia & Mauk, 1998; Krupa, Thompson, & Thompson, 1993; Lavond, Hembree, & Thompson, 1985; Steinmetz, Lavond, Ivkovich, Logan, & Thompson, 1992; Yeo, Hardiman, & Glickstein, 1985; but see also Welsh & Harvey, 1991). Because of some overlapping features of CR and CRM development, we recently investigated the role the IP may play in acquisition of CRM. Temporary inactivation of the IP with the GABA_A agonist muscimol during delay eyeblink conditioning blocked the acquisition of CRs and also the timing aspect of CRM, but did not prevent increases in the size of the UR (Burhans & Schreurs, 2018). When delay conditioning proceeded without inactivation, rabbits previously receiving muscimol were able to acquire CRs and also the increase in latency of the UR. These findings suggested that the IP plays a crucial role in the development of conditioning-related changes in the timing of the UR but not in the development of increases in the amplitude and area of the UR, the expression of which we have found to be dependent on the central nucleus of the amygdala (Burhans & Schreurs, 2008).

The following study was designed to investigate the role of the IP in the extinction of CRM using muscimol inactivation of the IP during unpaired extinction sessions following the establishment of eyeblink CRs and CRM. Although the contribution of the IP to extinction of eyeblink conditioning has not been studied as extensively as its role in acquisition, previous studies have shown that inactivation of the IP prevents expression of CRs during extinction and extinction learning itself, as revealed by a lack of extinction savings once IP inactivation is ceased (Hardiman, Ramnani, & Yeo, 1996; Ramnani & Yeo, 1996; Robleto & Thompson, 2008). Based on the previous literature and our recent findings, we hypothesized that inactivation of the IP during unpaired extinction would block expression and extinction of CRs and would prevent extinction of the latency aspect of CRM.

2. Methods

2.1. Subjects

The subjects were 18 male, New Zealand White rabbits (Oryctolagus cuniculus), 2–3 months of age, weighing approximately 1.8–2.3 kg upon delivery from the supplier (Charles River, Saint-Constant, Canada). The rabbits were housed in individual cages on a 12 hour light-dark cycle and given ad libitum access to food and water. They were maintained in accordance with the Guide for the Care and Use of Laboratory Animals issued by the National Institutes of Health, and the research was approved by the West Virginia University Animal Care and Use Committee. The experiment was run in four replications, with each run having 4–6 rabbits.

2.2. Surgical Procedure

After a minimum period of six days for adaptation to the animal facility, rabbits underwent aseptic stereotaxic surgery for implantation of a chronic guide cannula targeted at the right IP, ipsilateral to the trained eye. The rabbits were anesthetized using a subcutaneous injection (0.6 mg/kg) of a mixture of ketamine HCl (83.3 mg/ml) and xylazine (16.7 mg/ml) with additional 0.25–0.5 ml injections of ketamine (100 mg/ml) given every 30 −60 minutes to maintain anesthesia. Rabbits also received a single subcutaneous dose of the long lasting analgesic buprenorphine-SR (0.1 mg/kg) prior to the start of surgery, in addition to localized injection of bupivicaine (≤ 2.0 mg/kg), distributed along the incision site. Rabbits were given supplemental oxygen (1–2 %), and vital signs were continually monitored by a surgical assistant throughout the procedure.

After aligning the skull with Lambda 1.5 mm lower than Bregma, three holes were drilled, two for skull screws and one directly above the right IP using the following coordinates in reference to Lambda: AP= −0.5 AP (anterior to Lambda) and ML= −5.1, based on the atlas of Lavond and Steinmetz (2003). The holes for skull screws were placed on opposite sides of the skull, anterior to the hole above IP, and were fitted with small stainless steel screws (Small Parts Inc., Logansport, IN) that were partially screwed into the skull. To minimize the chance of brain hemorrhage, drilling above IP was done slowly and in stages, with frequent stops to place sterile bone wax into the hole to stop skull bleeding and improve visualization. A thin layer of skull was left intact and subsequently removed manually with a bone curette to avoid disturbing the dura and vessels below the skull. A stainless steel 22-gauge guide cannula, designed to custom fit an internal 28-gauge injection cannula with a 1.5 mm projection (Plastics One Inc., Roanoke, VA), was lowered slowly (DV= −13.0 from Lambda skull surface). Sterile bone wax was used to fill any empty space within the hole surrounding the cannula, and the cannula was then secured by anchoring it to the skull screws with dental acrylic. The guide was fitted with a protective dummy cannula of the same length that was removed for infusions. Following suturing, the surgical incision site was treated with triple antibiotic ointment, with continued application for three consecutive days thereafter.

2.3. Apparatus

The apparatus and recording procedures for eyeblink conditioning were first detailed by Schreurs and Alkon (1990) who modeled their apparatus based on those of Gormezano (Coleman & Gormezano, 1971; Gormezano, 1966). Briefly, rabbits were restrained in a Plexiglas box placed inside a sound-attenuating, ventilated chamber (Coulborn Instruments, Allentown, PA; Model E10–20). Inside the chamber, a stimulus panel containing a speaker and houselight (10-W, 120 V) was mounted at a 45° angle 15 cm anterior and dorsal to the rabbit’s head. An exhaust fan created a constant ambient noise level of 75 dB inside the chamber. Periorbital electrical stimulation was delivered by a programmable two-pole stimulator (Colbourn Instruments, Model E13–35) via stainless steel Autoclip wound clips (Stoelting, Wood Dale, IL) that were positioned 10 mm ventral and 10 mm posterior to the dorsal canthus of the right eye. Stimulus delivery, data collection, and analysis were all accomplished using the LabVIEW software system (National Instruments, Austin, TX).

As a measurement of the eyeblink, nictitating membrane responses (NMRs) were transduced by a potentiometer (Novotechnik US Inc., Southborough, MA; Model P2201) connected at one end, via a freely moving ball and socket joint, to an L-shaped lever containing a hook that attached to a 6–0 nylon loop that was sutured into but not through the nictitating membrane (NM) of the right eye. At the other end, the potentiometer was connected to a 12-bit analog-to-digital converter (5-ms sampling rate, 0.05-mm resolution), and individual A/D outputs were stored on a trial-by-trial basis for subsequent analysis.

2.4. Procedure

After at least six days for recovery from surgery, rabbits were first acclimated to restraint by being placed in restrainers for 30 minutes while under close supervision. Rabbits then received one training session per day in the following order: adaption, US pretest, six sessions classical delay eyeblink conditioning, US posttest (Post1), six sessions unpaired extinction with weak shock following muscimol or saline infusion into the IP, a second US posttest (Post2), and a CS-alone retention test (CS Test). There was a 48-hour break between the last session of extinction and Post2 to allow sufficient time for clearance of muscimol from the brain. Comparisons of Post1 with Pretest served as the initial evaluation of CRM following conditioning, while comparisons with Post2 measured CRM after extinction treatment with or without IP inactivation. The CS Test was used to evaluate extinction learning, savings, and retention.

For adaptation, subjects were restrained and prepared for delivery of the periorbital shock US and NMR recording and then adapted to the training chambers for an amount of time equivalent to subsequent training sessions (80 min). For pretest and posttests, subjects received 80 trials of US presentations with an average inter-trial interval (ITI) of 60 s (range 50–70 s). Each US presentation was one of 20 combinations of periorbital shock intensity (0.1, 0.3, 0.5, 1.0, or 2.0 mA) and duration (10, 25, 50, or 100 ms), and these 20 unique USs were presented in four separately randomized blocks with the restriction that the same intensity or duration could not occur more than three times in succession. For delay eyeblink conditioning, each session consisted of 80 trials of paired presentations of a 400 ms, 1 kHz, 82 dB CS that coterminated with a 100 ms, 2 mA US (300 ms interstimulus interval). The CS-US presentations were presented with an average ITI of 60 s. Unpaired extinction consisted of 80 presentations of the tone CS and 80 presentations of a reduced intensity 0.3 mA US (100 ms) that were explicitly unpaired and presented in a pseudorandom order such that the same stimulus was not presented more than three times in succession. To maintain the session length at approximately 80 minutes, the average ITI for unpaired sessions was reduced to 30 s. The CS test consisted of 80 presentations of the tone CS with an average ITI of 60 s.

2.5. Temporary Inactivation of the Interpositus Nucleus

One hour before the start of each of six days of extinction, rabbits were restrained and either the GABA_A agonist muscimol (Tocris Bioscience) dissolved in 0.9% saline (pH 7.4) or vehicle (0.9 % saline, pH 7.4) was infused into the right IP, ipsilateral to the trained eye. For all infusions, the dummy cannula was first removed, and 0.50 μl of muscimol (1 μg/μl,) or vehicle was infused at a rate of 0.25 μl /min through an injection cannula that when inserted, projected 1.5 mm from the tip of the guide cannula. Each injection cannula was attached to a 10 μl Hamilton microsyringe via polyethylene tubing, and the rate of infusion was controlled by an infusion pump (KD Scientific, Holliston, MA) that was modified to allow multiple rabbits to be infused simultaneously. Following infusion, the injection cannula remained in place for three minutes and was then replaced with the dummy cannula. Rabbits were returned to transport containers and monitored until they were prepared for behavioral training, which took place approximately one hour post-infusion.

2.6. Histology

After the completion of training, rabbits were anesthetized with a solution containing ketamine HCl (83.3 mg/ml) and xylazine (16.7 mg/ml), followed by a lethal dose of Euthasol (sodium pentobarbital, 390 mg/ml) and transcardial perfusion with 0.9% saline followed by 10% formalin. The brains were post-fixed in 10% formalin and prior to sectioning, cryoprotected in a 15% sucrose, 2% formalin solution for a minimum of 24 hours. Coronal sections were cut at 40 μm using a freezing microtome. Sections were placed on slides and digitally imaged while wet (1.5X) with an Olympus BX51 microscope equipped with a computer assisted camera and software (monochrome Olympus FVII, cellSens Standard v1.13). The digital images were then used to determine the placement of the guide cannula, verified by consulting two rabbit brain atlases (Lavond & Steinmetz, 2003; McBride & Klemm, 1968).

To further characterize the placement of guide cannula, a subset of rabbits (n=4) received infusions of fluorescently labeled muscimol (BODIPY TMR-X muscimol conjugate, Molecular Probes; 1 μg/μl dissolved in 0.5% dimethyl sulfate, 95% saline) at the same volume and rate as infusions during extinction training. After a minimum of one hour following infusion, rabbits were then perfused and tissue was processed and imaged as described above, with additional images taken to capture muscimol fluorescence (4X, U-MNG2 filter).

2.6. Data Analysis

For delay conditioning, CRs of the NMR were defined as any extension of the NM exceeding 0.5 mm that was initiated following CS onset but prior to US onset (i.e. the 300 ms ISI). For CS-alone presentations during extinction and the CS Test, the time window for CR initiation was extended to the first 600 ms post-CS in order to capture any longer latency CRs coinciding with the time frame when the UR to the US would have been initiated during paired CS-US presentations. For analysis of CRM, an unconditioned response (UR) during pre and post-testing was defined as any extension of the NM exceeding 0.5 mm that was initiated within 300 ms following US onset. The time window for measuring associative changes in the UR was based on prior observations that responses to the US following CS-US pairings had onset latencies within the same range as CRs (Schreurs et al., 2000) and follows the convention of two decades of CRM studies (Burhans et al., 2015, 2018; Burhans et al., 2008; Schreurs, 2003; Schreurs et al., 2011). URs to US-alone presentations during extinction were examined from US onset to the end of trial (1.5 sec).

Additional response measures included amplitude, area, and latency. Amplitude of the CR or UR was calculated as the maximum extension of the NM in millimeters. Onset latency of the response was the latency in ms from stimulus onset to when the NM rose 0.1 mm above baseline while peak latency was the latency in ms from stimulus onset until maximum NM extension occurred. Criterion latency was also calculated for CRs, defined as the latency in ms from CS onset to when the NM rose to the response criterion of 0.5 mm. Area of the response was calculated as the total area of the response curve (arbitrary units) from stimulus onset until the end of trial (1.8 sec for CRs and URs during US testing; 1.5 sec for URs during extinction). For URs during US testing, two additional measures were calculated in order to overcome the statistical limitations of empty data cells produced by subthreshold responses to periorbital shock, particularly at the lower intensities and durations. These measures, magnitude of the response amplitude (mAmp) and magnitude of the response area (mArea), included the amplitudes and areas of all NMRs above baseline regardless of whether the 0.5 mm criterion was met (Garcia, Mauk, Weidemann, & Kehoe, 2003).

A significant pre- to posttest increase in any of the UR response measures as a function of classical conditioning is a defining feature of CRM. To increase the sensitivity for detection of CRM and to follow the convention of previous CRM studies, data were collapsed at the five US intensities across duration and CRM analyses were focused on the first 20 trial US sequence where the strongest CRM is observed (Schreurs et al., 2000). To examine the shape and timing of NMRs during US tests, response topographies were generated at each US intensity by averaging across rabbits and across US durations within each experimental group.

Unless described otherwise, experimental group data were analyzed by repeated measures analysis of variance (ANOVA, SPSS 24), with p values corrected using the procedures of Huynh-Feldt for violations of the sphericity assumption. Planned and follow-up comparisons were Bonferroni corrected for the number of comparisons.

3. Results

3.1. Histology

One rabbit was not utilized in the study due to an adverse reaction to anesthesia during surgery, and a second rabbit was removed due to a failure to adapt to restraint during behavioral training. An additional rabbit was excluded from all analyses due to misplacement of the guide cannula, which was located in the medial wall of the ventral border of lobule HVI of the cerebellar cortex. Reconstructions of the location of the guide cannula tips for the remaining rabbits that received muscimol (black triangles, n=8) or saline (white circles, n=7) infusions into the IP are illustrated in Figure 1 (left panel). The majority of placements were typically in the dorsal portion of the IP or directly above, indicating that the injection cannula, which extended 1.5 mm beyond the guide tip, would likely have been within the IP when inserted. Also shown in Figure 1 (right panel) is an example coronal section from a muscimol rabbit that received an infusion of fluorescent muscimol one hour prior to perfusion. The corresponding fluorescent image illustrates a visible bolus of muscimol ventral to the tip with limited radial spread.

Figure 1.

Reconstruction of the location of guide cannula tips in the right interpositus nucleus for rabbits in the saline (white circles) and muscimol (black triangles) groups, shown in coronal drawings 1.5 to 0.5 mm anterior to Lambda. When inserted, injection cannula for infusions projected an additional 1.5 mm from the guide tip. On the top right is an example coronal brain section taken from a rabbit in the muscimol group that received infusion of fluorescent muscimol one hour prior to perfusion, with a fluorescent image of the boxed area shown below. FA: fastigial nucleus; DE: dentate nucleus; IP: interpositus nucleus.

3.2. Delay Eyeblink Conditioning

The average percentage of CRs (% CRs) to the tone CS across six sessions of delay eyeblink conditioning (D1–6) is shown on the left side of Figure 2, which also provides an overview of the experimental design along the X-axis. Rabbits rapidly acquired a high level of conditioning with an averaged final % CR of 98.9% (± 0.36 SEM) for both groups combined, as confirmed by a significant effect of Session [F(5,65) = 82.33, p < 0.001] with corrected planned comparisons indicating that CRs increased from the first to second session (p < 0.001) and remained at a similar high level across subsequent sessions. For analyses of % CRs and other CR measures (latency, amplitude, and area), there were no significant effects involving Group, demonstrating that both sets of rabbits were equivalent in learning rate and had similarly timed and sized CRs prior to IP infusions during extinction.

Figure 2.

Mean percentage (±SEM) of conditioned eyeblink responses (% CRs) to the tone conditioned stimulus (CS) during delay eyeblink conditioning (D1-D6), unpaired extinction with weak shock (E1–6), and a retention test of CS-alone presentations (CS Test) in rabbits receiving pre-extinction infusions of saline (white circles) or muscimol (black triangles) into the interpositus nucleus (IP). The X axis summarizes the experimental design and shows where unconditioned stimulus (US) testing sessions took place on days prior to (UR Pretest) and following delay (US Post1), and after extinction (US Post2).

3.3. Inactivation of the Interpositus during Unpaired Extinction

The average percentage of CRs to the tone CS across six sessions of unpaired extinction (E1–6) with weak shock is shown in the middle of Figure 2. In rabbits receiving muscimol infusions into the IP prior to extinction, expression of CRs was almost completely blocked, whereas rabbits receiving saline demonstrated a gradual reduction in CRs across sessions. This observation was confirmed by a significant interaction of Extinction Session by Group [F(5,65) = 10.06, p < 0.001], with corrected planned comparisons indicating that % CRs were significantly greater in saline compared to muscimol rabbits for E1 (p < 0.001), E2 and E3 (p’s < 0.01), E4 (p < 0.05), and E5 (trend, p = 0.072) but not for E6 (p = 0.222). Within group comparisons for saline rabbits indicated that responding decreased from E1 to E2 and from E2 to E3 (p’s < 0.05), and within-group paired t-tests comparing each extinction session with the last session of eyeblink conditioning (D6) demonstrated that saline rabbits showed moderately reduced % CRs at E1 (trend, p = 0.056) and significantly reduced responding of increasing magnitude throughout remaining sessions (E2: p < 0.05, E3–5: p’s < 0.01; E6: p < 0.001). Extinction sessions were also divided into four 20-trial blocks in order to examine within-session patterns of responding, shown in Figure 3. While saline rabbits demonstrated within-session decreases in responding indicative of extinction learning, the small percentage of CRs exhibited by muscimol rabbits did not consistently follow any within-session pattern. These observations were confirmed by a significant interaction of Block by Group [F(3,39) = 11.54, p < 0.001], with corrected post hoc comparisons demonstrating that only saline rabbits showed within-session decreases in responding, specifically from the first block to the second (p < 0.05), third (p < 0.01), and fourth blocks (p < 0.001).

Figure 3.

Mean percentage (±SEM) of conditioned eyeblink responses (% CRs) to the tone conditioned stimulus during six sessions of unpaired extinction with weak shock (E1–6), divided into blocks of 20 CS presentations (B1–4) in rabbits receiving pre-extinction infusions of saline (white circles) or muscimol (black triangles) into the interpositus nucleus.

The latency, amplitude, and area of the limited number of CRs expressed by muscimol rabbits was closely examined to determine if they had similar timing, size, and shape as CRs produced by saline rabbits throughout extinction. Because of the limitations of empty data cells due to the low and inconsistent level of responding in muscimol rabbits, analyses examined each extinction session separately for each CR measure. Data are shown for each measure in Figure 4, with statistics summarized in Table 1. For latency measures, CRs in muscimol rabbits had significantly greater criterion latencies compared to saline rabbits for all sessions of extinction and significantly greater onset and peak latencies for sessions E2–6. For amplitude and area measures, muscimol rabbits exhibited significantly smaller CRs compared to saline rabbits for the first three extinction sessions (E1–3). An example averaged CR topography for a saline and muscimol rabbit on the first day of extinction is also illustrated in Figure 4, further demonstrating that CRs produced under IP inactivation were smaller in size and slower to be initiated.

Figure 4.

Mean (±SEM) conditioned response (CR) onset latency (top left), peak latency (middle left), criterion latency (bottom left), amplitude (top right), and area (middle right) during delay eyeblink conditioning (D1–6) and unpaired extinction with weak shock (E1–6) in rabbits receiving pre-extinction infusions of saline (white circles) or muscimol (black triangles) into the interpositus nucleus. An example averaged CR topography for a saline (dotted line) and muscimol (solid black line) rabbit on the first day of extinction is shown on the bottom right.

Table 1.

Oneway ANOVA for Conditioned Eyeblink Response (CR) Measures during Extinction

	E1		E2		E3		E4		E5		E6
ONE-WAY ANOVA	F(1,12)	p	F(1,13)	p	F(1,10)	p	F(1,12)	p	F(1,12)	p	F(1,13)	p
CR Onset Latency	2.64	0.130	15.28	0.002	31.34	<0.001	92.58	<0.001	15.85	0.002	10.89	0.006
CR Peak Latency	3.52	0.085	31.29	<0.001	19.31	0.001	19.26	<0.001	6.63	0.024	12.98	0.003
CR Criterion Latency	5.22	0.041	40.19	<0.001	27.53	<0.001	62.71	<0.001	20.04	<0.001	11.86	0.004
CR Amplitude	10.08	0.008	9.17	0.010	9.10	0.013	1.12	0.310	0.64	0.439	0.226	0.643
CR Area	6.03	0.030	8.80	0.011	5.38	0.043	0.71	0.417	0.03	0.861	0.13	0.729

Note: Gray shading denotes p values ≤ 0.05

Similar to the pattern for % CRs, the amplitude and area of CRs in saline rabbits decreased across sessions but remained at a low, relatively fixed level in muscimol rabbits. Analyses focused on the saline group revealed significant effects of Extinction Session for amplitude [F(5,35) = 4.42, p < 0.001] and area [F(5,35) = 4.00, p = 0.025] with corrected post hoc comparisons indicating the size of the CR decreased from E1 to E6 for both amplitude (p < 0.01) and area (trend, p = 0.08).

To see if muscimol inactivation of the IP had any effects on the UR, eyeblink responses to the 0.3 mA shock-alone presentations during extinction were examined. Mean percent URs (% URs) averaged for the entire extinction session and also divided into 20-trial blocks are shown in Figure 5. There appeared to be an overall reduction in % URs for muscimol rabbits across extinction sessions, as confirmed by a main effect of Group [F(1,13) = 4.91, p = 0.045] with no significant interaction with Extinction Session. However, the block data revealed that although saline rabbits initially had higher levels of responding at the start of each session, both groups approached a similar terminal level by the end of each session. Analysis focused on comparisons of the first and last block indicated a significant interaction of Extinction Block by Group [F(1,13) = 7.63, p = 0.016] with corrected comparisons indicating that saline rabbits had greater % URs for the first block compared to muscimol rabbits (p< 0.001), but the difference was only a trend at the last block (p = 0.090). For within-group comparisons, both groups demonstrated the same pattern of decreased within-session responding from the first to the last block (p’s < 0.001). Similar to findings for CRs, the URs of muscimol rabbits overall had longer latencies, as indicated by a main effect of Group for onset latency [F(1,13) = 20.50, p = 0.001], peak latency [F(1,13) = 24.06, p < 0.001], and criterion latency [F(1,13) = 20.28, p = 0.001]. There was also some indication that they were smaller in size, with a trend for a main effect of Group for amplitude [F(1,13) = 3.42, p = 0.087], but there were no group effects for area.

Figure 5.

Mean percentage (±SEM) of unconditioned eyeblink responses (% URs) to the 0.3 mA unconditioned stimulus during six sessions of unpaired extinction with weak shock averaged for each session (E1–6) and divided into blocks of 20 US presentations (B1–4) in rabbits receiving pre-extinction infusions of saline (white circles) or muscimol (black triangles) into the interpositus nucleus.

3.4. Extinction Savings and Retention of CRs

Analysis of responding to the CS-alone presentations during the CS Test was used as a test for savings in muscimol rabbits to determine if extinction learning occurred despite the almost complete block of CR expression during extinction. To get a clear picture of % CRs at the very start of the session before the confounding effects of repeated CS presentations produced additional extinction, the CS Test was divided into 10-trial blocks, as shown in Figure 6. The averaged % CRs across the entire session can also be found on the right side of Figure 2, which allows comparisons of % CRs across the other phases of training. Based on the data averaged for the entire CS Test, it appears that muscimol rabbits did show savings, indicating that extinction learning took place and may have even been superior to saline rabbits. Examination of the data divided into blocks showed that muscimol and saline rabbits initially had a similar level of extinction retention, suggesting that the overall lower level of % CRs in muscimol rabbits was the result of a more rapid within-session decrease in responding. Analyses of the entire session averaged and also divided into blocks did not reveal any effects involving Group, indicating that rabbits with IP inactivation during extinction behaved as though they had acquired extinction learning at the same level as saline rabbits. The high amount of variability in responding during the CS Test in both groups may have occluded the ability to detect group differences for within-session responding across blocks. Analyses of the latency, amplitude, and area of the CRs produced during the CS Test also did not reveal any effects involving Group, indicating that CRs in muscimol rabbits had a similar timing, shape, and size as saline rabbits.

Figure 6.

Mean percentage (±SEM) of conditioned eyeblink responses (% CRs) to the tone conditioned stimulus during the retention test with conditioned stimulus-alone presentations (CS Test) following extinction, divided into blocks of 10 trials (B1–8) in rabbits previously receiving pre-extinction infusions of saline (white circles) or muscimol (black triangles) into the interpositus nucleus.

Because muscimol rabbits exhibited low levels of CRs during extinction, ranging from 2.8 to 12.8%, individual subject data were examined to see if there were any patterns between extinction CR levels and responding during the CS Test. The muscimol rabbit with the highest level of CRs during extinction had the lowest level of responding during the CS Test (3.9%), whereas the rabbit with the lowest level of responding during extinction had the highest level of responding during the CS Test (87.5%). Although these two extremes suggest that extinction learning was more likely to occur if CRs were expressed during extinction, the remaining rabbits did not follow as clear a pattern. In addition, the range of responding did not relate to differences in the placement of the cannula within IP. Consequently, Pearson’s correlation analysis did not reveal a significant relationship between averaged % CRs during extinction and the CS Test (r = −0.36, p = 0.43) in muscimol rabbits, but the negative correlation was stronger when the comparison was restricted to the first 10 trial block of the CS Test (r = −0.69, p = 0.14), a more sensitive measure of extinction retention. It should also be noted that the range of responding during the CS Test was quite variable in saline rabbits as well, ranging from 2.5 to 98.7 % CRs. However, for saline rabbits, there was a significant positive correlation between responding during the last session of extinction and the CS Test, averaged across the entire session (r = 0.82, p = 0.013) and for the first 10 trials of the CS Test (r = 0.86, p = 0.006), indicating that responding during the CS Test was a reflection of the terminal level of CRs during extinction. In other words, rabbits that had higher levels of responding during the last extinction session tended to have higher levels of responding during the CS Test.

3.5. Conditioning-Specific Reflex Modification

Comparisons of Pretest to the first posttest following eyeblink conditioning (Post1) were conducted to determine the initial level of CRM prior to extinction training with IP infusions. Indicative of CRM, Pretest to Post1 increases in UR latency measures (onset and peak), magnitude of the amplitude (mAmp) and area (mArea), and mean % URs were observed, particularly at the intermediate US intensities, 0.3 mA and 0.5 mA. Changes in the latency, amplitude, and area of the UR can be discerned by the UR topographies in Figure 7, while % URs, which cannot be discerned from topographies, are shown in Figure 8. For latency measures, analyses focused on the intensities for which a majority of rabbits responded (0.3 to 2.0 mA) to avoid limitations from empty data cells. There was a significant interaction of US Test by Intensity for onset latency [F(3,36) = 4.67, p < 0.05] and peak latency [F(3,36) = 7.52, p < 0.001]. Confirming significant CRM at intermediate US intensities, corrected planned comparisons demonstrated that a Pretest to Post1 increase occurred for onset latency at 0.3 mA (p < 0.01), 0.5 mA (trend, p = 0.064), and 1.0 mA (p < 0.05), and for peak latency at 0.3 mA (p < 0.05). CRM for the size and frequency of the UR was indicated by a significant interaction of US Test by Intensity for mAmp [F(4,52) = 8.14, p < 0.001], mArea [F(4,52) = 6.89, p < 0.001], and % URs [F(4,52) = 4.07, p< 0.01]. Although Pretest to Post1 increases were found for these measures at the 0.3 and 0.5 mA intensities, the corrected comparisons were not significant. There were no effects of Group for any UR measures, indicating that both groups demonstrated similar levels of CRM prior to extinction with IP infusions.

Figure 7.

Eyeblink unconditioned response topographies to unconditioned stimulus presentations prior to delay eyeblink conditioning (Pretest, dotted black line), following delay conditioning (Post1, black line), and following unpaired extinction with weak shock with pre-extinction infusions of saline or muscimol (Post2, gray line) into the interpositus nucleus. Data shown are for the first 20 trials of pre- and posttesting, averaged at each of five US intensities (0.1 to 2.0 mA) and collapsed across US duration.

Figure 8.

Mean percentage (±SEM) of unconditioned responses (% URs) to unconditioned stimulus presentations prior to delay eyeblink conditioning (Pretest, white bar), following delay conditioning (Post1, black bar), and following unpaired extinction with weak shock with pre-extinction infusions of saline or muscimol (Post2, gray bar) into the interpositus nucleus. Data shown are for the first 20 trials of pre- and posttesting, averaged at each of five US intensities (0.1 to 2.0 mA) and collapsed across US duration.

3.6. Extinction Effects on Conditioning-Specific Reflex Modification

The US posttest following extinction (Post2) was a test for the effects of extinction treatment under IP inactivation on extinction of CRM. Previous studies had established that six days of unpaired extinction with weak shock, as utilized in this study, produced significant decreases in the latency and size of CRM (Burhans et al., 2015, 2018; Schreurs et al., 2011, 2018). Examination of the UR topographies in Figure 7, however, suggested that increased size of the UR at intermediate intensities at Post1 was maintained and possibly potentiated at Post2 regardless of the state of the IP during extinction treatment. Latency measures, however, did appear to extinguish, returning to Pretest levels in both groups.

To specifically address extinction and IP inactivation effects on CRM, analyses of MAmp, MArea, and % URs focused on 0.3 and 0.5 mA, the intensities that initially showed CRM prior to extinction. For analysis comparing Post1 to Post 2 for MAmp, there was a trend for an interaction of US Test by Intensity by Group [F(1,13) = 3.83, p = 0.072]. Although corrected planned comparisons did not yield any significant differences, the main contributor to the interaction appeared to be the larger Post1 to Post2 increase in amplitude at 0.3 mA for the muscimol group. For analysis of MArea, there were no effects involving US Test or Group, reflecting no Post1 to Post2 change in area in either group. For % URs, there was an interaction of US Test by Intensity [F(1,13) = 7.02, p = 0.020] with no significant effects of Group. Although follow up comparisons were not significant, there was a pattern for greater Post1 to Post2 increases at 0.3 mA compared to 0.5 mA. Importantly, Post2 analyses comparing groups at each intensity (0.1– 2.0 mA) revealed no group differences, demonstrating that previous IP inactivation did not have any general effects on expression of the UR in terms of size and frequency.

Because CRM does not occur in all subjects and strong CRM only occurs in approximately 25% of subjects (Smith-Bell, Burhans, & Schreurs, 2012; Smith-Bell & Schreurs, 2017), individual UR topographies were examined to make sure averaging US Test data did not obscure any effects of IP inactivation. Although there were not enough rabbits to divide into groups based on CRM levels for quantitative analysis, both groups were found to have examples of rabbits with CRM that extinguished after extinction treatment but also those that showed extinction resistance. The variability in response to extinction treatment did not seem to be affected by IP inactivation during extinction. In addition, CRM extinction did not appear to relate to CR expression levels during extinction in muscimol rabbits.

4. Discussion

The main goal of the current study was to investigate the role of the IP in the extinction of CRM, associative changes in the UR that are characterized by increases in the latency, amplitude, area, and frequency of URs to reduced intensity US-alone presentations following eyeblink conditioning (Schreurs et al., 1995). We hypothesized that temporary inactivation of the IP with muscimol during unpaired extinction, an extinction procedure previously shown to reduce both eyeblink CRs and CRM (Schreurs et al., 2000; Schreurs et al., 2011), would prevent extinction of the latency aspect of CRM, the acquisition of which was recently shown to be dependent on the IP (Burhans & Schreurs, 2018). In addition, it was expected that IP inactivation would block the expression and extinction of CRs based on other reports in the eyeblink conditioning literature (e.g. Ramnani & Yeo, 1996; Robleto & Thompson, 2008). Concerning CRM, the major finding was that the inactivation state of the IP during extinction did not have any significant effects on extinction of CRM timing, as both muscimol and control rabbits showed reductions in UR latency measures. The second major finding was that despite the almost complete blockade of CR expression during extinction under IP inactivation, extinction learning did appear to take place, as demonstrated by significant extinction savings once IP inactivation was removed. Another interesting result was the observation of low levels of atypical small, long-latency CRs in rabbits with IP inactivation during extinction.

4.1. Possible cerebellar mechanisms for associative timing changes in the UR

Our recent work established that plasticity within the IP is critical for the acquisition of the associative change in the timing of the UR (Burhans & Schreurs, 2018). More specifically, rabbits with muscimol inactivation of the IP during delay conditioning did not show increases in the latency of the UR when tested for CRM but were able to develop those changes once conditioning proceeded without inactivation. The increase in peak latency of the UR that occurs with CRM is very reminiscent of the timing of the CR during CS-US pairings, where the peak of the response moves forward, approaching the time when the onset of the US would have occurred. We therefore theorized that a similar mechanism controlling the timing of the CR may also control the timing of the UR to low intensity US-alone presentations following conditioning. For CRs, this precise timing is thought to reflect the development of plasticity within the cerebellar cortex and IP. More specifically, CS-US pairings induces long-term depression at cortical Purkinje cell (PC) synapses where CS and US information converges (Freeman, Shi, & Schreurs, 1998; Ito & Kano, 1982; Linden & Connor, 1991; Schreurs, Oh, & Alkon, 1996; but see also Welsh et al., 2005), resulting in a precisely timed firing pause that disinhibits the IP, allowing the CR to be initiated ((De Zeeuw & Berrebi, 1995; Halverson, Khilkevich, & Mauk, 2015; Heiney, Kim, Augustine, & Medina, 2014; Ito, Yoshida, Obata, Kawai, & Udo, 1970; Jirenhed et al., 2007; Rasmussen, Jirenhed, Wetmore, & Hesslow, 2014). One possibility then is that the plasticity required within the IP to induce timing changes in the UR observed for CRM is initiated by cortical inputs to the IP.

The current study has confirmed, however, that IP is not needed to extinguish CRM timing, and it’s important to note that all aspects of CRM, including the latency increase, do not simply fade with the passage of time, persisting for at least a month following conditioning (Schreurs et al., 2018). Therefore, the extinction of CRM is likely an active process with distinct neural substrates. Fitting our findings that IP plasticity is required for CRM acquisition but not extinction, extensive work by Medina and colleagues has suggested a mechanism whereby extinction induces new plasticity in the cortex while the acquisition-induced plasticity in the IP can persist (for review, see Mauk et al., 2014). Therefore, a prime candidate for extinction of CRM timing is the cerebellar cortex, which we have previously argued may also play an important role in the acquisition of CRM (Burhans & Schreurs, 2018). Important future directions would be to explicitly examine the role of plasticity within the cerebellar cortex for both the acquisition and extinction of CRM.

Of note, in this study we did not find a significant reduction in the size of CRM following extinction treatment in muscimol rabbits nor in saline controls, despite several previous publications demonstrating that six days of unpaired extinction with weak shock induces significant decreases in amplitude and area of CRM (Burhans et al., 2015, 2018; Schreurs et al., 2011, 2018). Because we typically see a fair degree of variability in terms of individual responsiveness to extinction treatment, it may be that by chance, we had more extinction-resistant subjects in this cohort. However, another possibility is that the IP cannula track damage to the overlaying cerebellar cortex, which would have occurred for both saline and muscimol groups, may have affected CRM extinction. Work by others has shown that rabbits with post-training unilateral cerebellar cortical lesions that include regions directly above the deep cerebellar nuclei (i.e. lobule HVI) show increased amplitude URs to US-alone presentations across a range of intensities, suggesting there may be cortical modulation of CRM-like changes in UR amplitude (Gruart & Yeo, 1995). The seeming lack of extinction of the area and amplitude of CRM may also be the results of a renewal effect (Bouton, 2002; Grillon, Alvarez, Johnson, & Chavis, 2008), as the infusion procedure taking place prior to extinction sessions may have constituted a change in context from CRM testing where no infusions take place. We have found previously that expression of CRM is sensitive to a change in context between acquisition and CRM testing if both contexts are familiar (Schreurs, Gonzalez-Joekes, & Smith-Bell, 2006), so it is feasible that at least some aspects of CRM extinction may be context-dependent as well.

4.2. Extinction savings after IP inactivation- Alternative sites for extinction plasticity

One of the unexpected findings of the current study was the evidence for extinction savings during CS-alone testing in rabbits with previous IP inactivation during extinction. The volume, timing, and stereotaxic coordinates of muscimol infusions in our study were partly modeled on those utilized by Thompson and colleagues in several publications on muscimol inactivation of IP during acquisition or extinction of delay eyeblink conditioning in rabbits (Krupa et al., 1993; Krupa & Thompson, 1997; Robleto & Thompson, 2008). In their 2008 paper investigating extinction, muscimol inactivation of the IP prior to four daily sessions of CS-alone extinction blocked expression of CRs and extinction learning with no evidence for savings (Robleto & Thompson, 2008), only the former of which we replicated in the current study. Work by others has also clearly demonstrated a lack of extinction savings after muscimol inactivation of the IP during CS-alone extinction (Hardiman et al., 1996; Kalmbach & Mauk, 2012; Ramnani & Yeo, 1996). Of significance, our previous study examining IP inactivation during eyeblink conditioning blocked both the acquisition and expression of CRs, demonstrating that our inactivation procedure, the same used in the current study, is capable of blocking plasticity within the IP (Burhans & Schreurs, 2018).

The most obvious distinction between our study and many others in the literature is that instead of CS-alone presentations, we used unpaired extinction with weak shock. We chose the unpaired extinction protocol because it was previously established as the most optimal treatment to simultaneously reduce both CRs and CRM (Burhans et al., 2015; Schreurs et al., 2000). A point of emphasis of our previous work on the CRM paradigm has been to validate it as a corollary to current animal models of post-traumatic stress disorder (PTSD), as we have argued extensively that CRM is a form of conditioned hyperarousal that shares many commonalities with the hypervigilance symptoms of PTSD (Burhans et al., 2008; Schreurs, 2003; Schreurs & Burhans, 2015), just one aspect of the complex symptomology of PTSD (American Psychiatric Association, 2013). Of importance to note, the clinical use of US-alone presentations during extinction has gained interest, as it has shown superiority to traditional CS-alone presentations in terms of its ability to thwart fear renewal and spontaneous recovery in animal and human studies (Leer, Haesen, & Vervliet, 2018; Rauhut, Thomas, & Ayres, 2001; Thompson, McEvoy, & Lipp, 2018; Vervliet, Vansteenwegen, & Hermans, 2010). In order to find treatments that were more tenable for human clinical use, we examined whether reducing the intensity of the US presented during unpaired extinction would still be successful at reducing CRs and CRM (Schreurs et al., 2011). As this was found to be the case, we adopted unpaired extinction with weak shock as our standard CRM extinction protocol for all future experiments. In our current work, despite the almost complete block of CRs with IP inactivation during extinction, rabbits continued to exhibit URs to US-alone presentations, albeit less frequently and with a longer latency than controls. This persistence of eyeblink responding throughout extinction may have compensated for the IP blockade, perhaps by promoting extinction-related plasticity in alternative areas.

Whether or not responding is necessary during extinction for extinction learning to occur is an interesting question. Earlier work by Thompson and colleagues showed that inactivation of the motor nuclei during extinction of eyeblink conditioning blocked the expression of CRs but also extinction learning, suggesting that response expression was a requirement (Krupa & Thompson, 2003). Follow up work examined the relationship between response expression and extinction learning by combining IP inactivation with precisely timed stimulation of the red nucleus (RN) to elicit eyeblinks following the presentation of the CS, in addition to a separate experiment that examined the effects of the inactivation of the RN by itself. Results showed that extinction learning occurred with both manipulations despite a large difference in response expression, being very high for combined IP inactivation-red nucleus stimulation and very low for RN inactivation. One interpretation of these findings that may also explain our results is that the RN, rather than being just a simple relay for the CR, may be capable of extinction-related plasticity in the absence of IP input, either by direct stimulation or if stimulated by US-alone presentations. The ability of the RN to exhibit learning-related activity when inputs are altered has also been shown for acquisition of delay conditioning when inputs from motor cortex are blocked (Pacheco-Calderon, Carretero-Guillen, Delgado-Garcia, & Gruart, 2012).

An alternative mechanism that may be allowing extinction learning to occur in the absence of the IP could involve the inferior olive (IO), which relays US input to both the cerebellar cortex and IP via climbing fibers (Brodal, Walberg, & Hoddevik, 1975; Sugihara, Wu, & Shinoda, 1996; Van Ham & Yeo, 1992). The IO has been described as a reinforcement or error detection signal that drives learning-related changes in cortical PCs, and this signal is in turn subject to inhibitory feedback from the IP (for review, see Freeman & Steinmetz, 2011; Rasmussen & Hesslow, 2014; Robleto, Poulos, & Thompson, 2004). During acquisition of delay conditioning, the IO initially exhibits US-induced activity that then wanes as CRs emerge (Hesslow & Ivarsson, 1996; Kim, Krupa, & Thompson, 1998; Nicholson & Freeman, 2003; Sears & Steinmetz, 1991). When CS-alone presentations occur during extinction, the IP initially maintains that inhibitory signal as CRs continue to be expressed, but the complete absence of the excitatory US pushes the IO activity to below baseline levels. The resulting change in the pattern of climbing fiber input to the PCs is thought to serve as a signal for extinction, supported by studies showing that IO inhibition leads to extinction-like decreases in CRs even with continued CS-US pairings (Bengtsson, Jirenhed, Svensson, & Hesslow, 2007; McCormick, Steinmetz, & Thompson, 1985; Medina, Nores, & Mauk, 2002; but see also Zbarska, Bloedel, & Bracha, 2008). However, in the current study, US-alone presentations during extinction would continue to activate climbing fiber input to the PCs that is importantly not timed with CS input, serving as an alternative extinction signal.

4.3. CRs expressed during IP inactivation

Analysis of the low levels of CRs performed in rabbits with IP inactivation during extinction revealed CRs with significantly longer onset, criterion, and peak latencies that were smaller in size, for both amplitude and area measures. A simple explanation for these atypical, infrequent CRs is that they represent spontaneous blinks, which may be perpetuated by the presence of the US throughout extinction, or breakthrough responding of CRs due to incomplete IP inactivation and/or off-target effects from muscimol spread to other areas. An alternative idea is that the CRs observed during IP inactivation may reflect eyeblink-related plasticity in extra-cerebellar sites that can exert some control over CRs (Ammann et al., 2016; Pacheco-Calderon et al., 2012). Another interesting possibility is that the CRs may be generated by the contralateral cerebellum, as we have previously demonstrated bilateral labeling in the deep cerebellar nuclei following unilateral retrograde viral injection into the rabbit eyelid (Gonzalez-Joekes & Schreurs, 2012). CRs as well as URs in the eye contralateral to the application of the US have been observed across species (e.g. Campolattaro & Freeman, 2009; Hilgard & Marquis, 1936; Ivarsson & Hesslow, 1993), and for rabbits in particular are described as infrequent and smaller in size with longer latencies, similar to the CRs observed here (Disterhoft, Kwan, & Lo, 1977; Lee, Kim, & Wagner, 2008; McCormick, Lavond, & Thompson, 1982).

4.4. Conclusions

Findings presented here show that the IP is not necessary for the extinction of eyeblink CRs or CRM timing during unpaired extinction treatment, despite its crucial role in acquisition and CS-alone extinction, adding support to the idea that multiple cerebellar and extra-cerebellar sites are capable of supporting eyeblink conditioning and/or its extinction under varying conditions. This multi-site plasticity concept has gained more traction in the field (D’Angelo et al., 2016; Kalmbach & Mauk, 2012) and provides some unity to the contrasting findings on the mechanisms of eyeblink conditioning in the literature. In addition, multi-site plasticity may be important for explaining individual differences in acquisition and extinction rates for CRs and CRM (Halverson, Hoffmann, Kim, Kish, & Mauk, 2016; Kalmbach & Mauk, 2012; Smith-Bell et al., 2012; Smith-Bell & Schreurs, 2017). Perhaps most importantly, a distributed plasticity view allows room for flexibility and adaptability in eyeblink conditioning, an ideal feature of any model of associative learning.