Amygdala conditioning modulates sensory input to the cerebellum

Abstract

Localization of emotional learning in the amygdala and discrete motor learning in the cerebellum provides empirical means to study the mechanisms mediating the interaction between fast emotional and slow motor learning. Behavioral studies have demonstrated that fear conditioning facilitates the motor conditioning. The present study tests the hypothesis that the amygdala output induces this facilitation by increasing the salience of the conditioned stimulus (CS) representation in the pontine nucleus (PN) input to the cerebellum. Paired trials of CS–US (unconditioned stimulus) were applied to anesthetized rats, a condition that allows for amygdala-based fear conditioning but not cerebellar-based motor conditioning. Multiple unit recordings in the PN served to assess the salience of the CS. Results showed that CS–US conditioning increased the PN-reactivity to the CS. Lidocaine-induced reversible inactivation of the amygdala prevented the facilitatory effect of conditioning on the PN-reactivity to the CS. These findings suggest that the amygdala-based conditioned responses reach the PN and increase the salience of the CS signal there, perhaps facilitating cerebellar conditioning. This facilitatory effect of the amygdala may be conceptualized under the ‘two-stage theory of learning’, which predicts that emotional learning in the first stage accelerates the motor learning in the second stage. We hereby demonstrate the physiological mechanism through which fast emotional learning in the first stage facilitates slow cerebellar learning in the second stage.

Introduction

Early studies of delayed eyeblink conditioning demonstrated retaining of conditioned eyeblink responses in decerebrated rabbits (Mauk & Thompson, 1987). These findings support the essential and sufficient role of cerebellar plasticity in the classical delay eyeblink conditioning paradigm (Bracha et al., 1995, Bracha et al., 1994, Bracha et al., 2001, Christian and Thompson, 2003). Other studies demonstrated severely deficient conditioning after decerebration in rats, rabbits and cats (Hesslow, 1994, Lovick and Zbrozyna, 1975, Mauk and Thompson, 1987), although normal conditioning was achieved after increasing the intensity of the CS (Hesslow, 1994). These findings point to the role of the cerebrum in enhancing the salience of the sensory inputs to the cerebellum, thus enabling the realization of the cerebellar plasticity. As a consequence, although the cerebellum is an essential locus of plasticity for eyeblink conditioning, specific constraints may prevent it from acquiring motor conditioned responses. The challenge set by this dogma is to map the extra-cerebellar brain sites that contribute to the conditioning process and the mechanisms through which they do so (Berger and Thompson, 1978, Christian and Thompson, 2003, Maren, 1999, Maren et al., 1997, Steinmetz, 2000).

A relevant example is a study in which lesioning of the rat’s amygdala impaired the cerebellum-based eyeblink conditioning to tone-CS of 65 but not 85 dB (Weisz, Harden, & Xiang, 1992). A related finding was the absence of eyeblink conditioning to tone-CS but a normal conditioning to an electrical-train stimulation in the PN in 12 day old pup rats (Campolattaro & Freeman, 2008). Given that the PN is the last pre-cerebellar site for convergence of CS-related afferents, the authors concluded that the “cerebellar neurons are capable of supporting associative learning… if given sufficient afferent stimulation”. Based on these studies, we suggest that: (i) the cerebellum comprises the essential and sufficient site of plasticity for motor conditioning (Mauk & Thompson, 1987), (ii) cerebellar plasticity can be realized by a salient CS-signal at the PN-input to the cerebellum, resulting in motor conditioning: an example is acceleration of eyeblink conditioning with an increase in intensity of the tone-CS (Campolattaro and Freeman, 2008, Weisz et al., 1992), and (iii) cerebellar plasticity cannot be realized by a non-salient CS-signal in the PN-input to the cerebellum. Hence, contribution of the extra-cerebellar sites to motor conditioning may be particularly necessary when a low-intensity CS is applied. The mechanism through which the cerebrum facilitates the cerebellar motor conditioning is currently not known.

The amygdala is a feasible candidate for facilitating the salience of the CS-signal in the PN-input to the cerebellum (Neufeld and Mintz, 2001, Thompson et al., 1987, Weisz et al., 1992). The amygdala is known for its fast-acquisition of fear-CRs, which can be achieved after just a few CS–US trials (Berger and Thompson, 1978, Davis, 1992, LeDoux et al., 1990, Sears et al., 1996, Steinmetz, 2000). The acquired neuronal-CRs may be projected from the central nucleus of the amygdala (CeA) through monosynaptic or polysynaptic pathways to the PN (Berger et al., 1976, Berger et al., 1986, Kandler and Herbert, 1991, Mihailoff et al., 1989), enhancing the PN-reactivity to the CS-signal arriving through the auditory–sensory projections. The final result of this process would be increased saliency of the CS-signal in the PN-input to the cerebellum, which in turn will accelerate the cerebellar learning process.

The above argumentation can be conveniently conceptualized under the ‘two-stage theory of learning’ (Konorski, 1967, May, 1948, Overmeir and Lowry, 1979, Prokasy, 1972, Rescorla and Solomon, 1967, Thompson et al., 1987, Weinberger, 1982, Weisz et al., 1992). It predicts that the fast-acquisition of limbic-based emotional-CRs during the first stage of learning facilitates the slower acquisition of cerebellum-based motor-CRs during the second stage of learning (Lennartz and Weinberger, 1992, Mintz and Wang-Ninio, 2001, Neufeld and Mintz, 2001, Thompson et al., 1987). Accordingly, amygdala lesions impaired the acquisition of eyeblink-CRs (Beylin et al., 2001, Blankenship et al., 2005, Lee and Kim, 2004, Mintz and Wang-Ninio, 2001, Sanders et al., 2003, Weisz et al., 1992), and amygdala-based fear conditioning facilitated the subsequent eyeblink conditioning to the same tone-CS (Neufeld & Mintz, 2001).

In the present study we tested whether CS–US conditioning in the amygdala enhances the salience of the CS-signal in the PN-input to the cerebellum. Study 1 demonstrates that CS–US conditioning increased the PN-reactivity to low-intensity tone-CS. Study 2 demonstrates that reversible inactivation of the amygdala abolished the conditioning-related increase in the PN-reactivity to tone-CS.