Behavioral and in vitro correlates of compulsive-like food seeking induced by operant conditioning in Aplysia

Abstract

Motivated behaviors comprise appetitive actions whose occurrence results partly from an internally driven incentive to act. Such impulsive behavior can also be regulated by external rewarding stimuli that, through learning processes, can lead to accelerated and seemingly automatic, compulsive-like recurrences of the rewarded act. Here, we explored such behavioral plasticity in Aplysia by analyzing how appetitive reward stimulation in a form of operant conditioning can modify a goal-directed component of the animal's food-seeking behavior. In naive animals, protraction/retraction cycles of the tongue-like radula are expressed sporadically with highly variable interbite intervals. In contrast, animals that were previously given a food-reward stimulus in association with each spontaneous radula bite now expressed movement cycles with an elevated frequency and a stereotyped rhythmic organization. This rate increase and regularization, which was retained for several hours after training, depended on both the reward quality and its contingency because accelerated, stereotyped biting was not induced in animals that had previously received a less-palatable food stimulus or had been subjected to nonassociative reward stimulation. Neuronal correlates of these learning-induced changes were also expressed in the radula motor pattern-generating circuitry of isolated buccal ganglia. In such in vitro preparations, moreover, manipulation of the burst frequency of the bilateral motor pattern-initiating B63 interneurons indicated that the regularization of radula motor pattern generation in contingently trained animals occurred separately from an increase in cycle rate, thereby suggesting independent processes of network plasticity. These data therefore suggest that operant conditioning can induce compulsive-like actions in Aplysia feeding behavior and provide a substrate for a cellular analysis of the underlying mechanisms.

Figure 1.

Behavioral protocol. A, During food-seeking behavior induced by continuous application of a seaweed stimulus to the lips, the radula of Aplysia produces cycles of biting movement consisting of a protraction (frames 2, 3), then a retraction phase (frame 4). Onset of a bite cycle was taken as the instant at which the radula became visible (frame 2; see vertical bar). B, Biting movements were analyzed during two pretraining and posttraining test periods that each lasted 20 min or until 100 successive radula cycles had occurred within a maximum sample time of 40 min. Immediately after the first test period, a 40 min training protocol was applied to different experimental groups of animals as described in Materials and Methods. After training, animals were rested (without any food stimulus) for 1–24 h, and then biting movements were again analyzed under the same experimental conditions as for the pretraining test period.

Figure 2.

Food-reward training induces a long-lasting increase in radula cycle rate. A, B, Sample recordings of successive bites (vertical bars) in a naive animal (A) and 1 h after food-stimulus training in three different animals (B), one of which received the inciting stimulus alone (without reward stimulation) during the training period (Control; 1) or had received (and ingested) a piece of seaweed (Seaweed; 2) or an equivalent sized patch of cellulose (Cellulose; 3) in association with each spontaneous radula protraction. C, Group analysis of behavioral changes. The frequency of radula biting in the posttraining test period was significantly higher in the seaweed-rewarded group (unfilled bar; see also 2) compared with the control (black bar; q₂ = 4.704) or cellulose-rewarded groups (gray bar; q₃ = 4.167). The cellulose and control groups were not significantly different [not significant (n.s.); q₂ = 1.497].

Figure 3.

Food-reward induced rhythmicity of radula biting movements. A, Autocorrelation analysis of 100 successive radula cycles 1 h posttraining in three different control (1), seaweed-trained (2), and cellulose-trained (3) animals. Top row is autocorrelation raster plots of intervals (dots) between onsets of successive bite cycles; bottom row is corresponding frequency histograms. Bite cycles occurred randomly in animals from the control and cellulose groups (indicated by flat histograms that were not fitted significantly by a Gabor function). In contrast, an animal from the seaweed group displayed regular periodic (∼5 s per cycle) recurrences of biting as indicated by a significant (p < 0.001) Gabor function fit of the data (bold line in A2, lower). B, Group comparison. The proportion of animals that expressed rhythmic biting was significantly increased in the seaweed group compared with either the control (p < 0.02) or cellulose (p < 0.02) groups. The latter two groups were not significantly different (p = 1).

Figure 4.

Experimental protocols for contingent/noncontingent food-reward training. A, Control animal group in which radula bite cycles (vertical bars) were monitored during inciting lip stimulation only. B, Contingent group in which a food-reward stimulus (20 μl of seaweed juice) was additionally injected (at arrowheads) into the buccal cavity in strict association with each spontaneous bite cycle. C, Noncontingent group in which seaweed juice was injected at 6 s intervals (at arrowheads) into the buccal cavity with no explicit association with radula movements. Note that animals in the contingent and noncontingent groups received the same number of rewarding stimuli per training period.

Figure 5.

Contingent-dependent increase in rate of radula biting. A, Sample recordings of successive onsets of bite cycles (vertical bars) during a test period at 1 h after training (according to protocols in Fig. 4) in a control (1), a contingently rewarded (2), and a noncontingently rewarded (3) animal. B, Group comparison. The rate of biting was higher after contingent-reward stimulation than in control or noncontingent animals [control vs contingent, q₃ = 4.750; contingent vs noncontingent, q₂ = 6.463; control vs noncontingent, q₂ = 0.626, not significant (n.s.)].

Figure 6.

Rhythmic biting induced by contingent food-reward stimulation. A, Autocorrelation analysis (top, raster plots; bottom, corresponding frequency histograms) of 100 successive radula cycles at 1 h after training. In different control (1) and noncontingent (3) animals, bites occurred randomly (histograms not fitted by a Gabor function). In contrast, an animal from the seaweed group (2) expressed rhythmic biting (period of ∼4.8 s) as indicated by the data fit (p < 0.001) with a Gabor function (bold line, bottom). B, Group comparison. The proportion of animals that expressed rhythmic biting was significantly higher in the contingent (Cont.) group than either the control (p < 0.001) or noncontingent (Non-cont.; p < 0.001) groups. Control versus noncontingent, not significant (n.s.).

Figure 7.

Persistence of the changes in radula behavior. A, B, Comparisons of radula cycle rates and rhythmicity in new control, contingent (Cont.), and noncontingent (Non-cont.) groups of animals tested 4 h (A) and 24 h (B) after training. As at 1 h after training (see Figs. 5B, 6B), after 4 h the rate of biting (A1; control vs contingent, q₃ = 4.500; noncontingent vs contingent, q₂ = 3.988; noncontingent vs control, q₂ = 2.659) and the proportions of animals expressing biting rhythmicity (A2) remained significantly higher (p < 0.05) in the contingent group than in either the control or noncontingent groups. By 24 h, however, significant intergroup differences in rate (B1) or rhythmicity of biting (B2) no longer occurred.

Figure 8.

Neuronal correlates of contingent-dependent behavioral changes. A1, Experimental protocol. After contingent, noncontingent, or control training (see also Fig. 4), the buccal ganglia of animals (n = 7) from each group were immediately isolated and the generation of radula motor output (fictive biting) was tested within 4 h of in vivo training. A2, Schematic of a bilateral buccal ganglia (B.g.) preparation from which extracellular recordings (indicated by black dots) were made from the I2n. (radula protractor motoneurons), n.2,1 (retractor motoneurons), and R n. (closure motoneurons). The B63 radula pattern-initiating neurons were also identified and recorded intrasomatically (arrowheads) in the two buccal ganglia. During in vitro testing, a continuous inciting stimulus was provided by monotonic electrical stimulation (2 Hz, 8.5 V) of the bilateral n.2,3 sensory nerves. B–D, Samples of simultaneous extracellular recordings of radula motor output from indicated motor nerves (I2n., n.2,1., R n.) and intracellular recordings from the left (l) and right (r) B63 pattern-generating neurons during the posttraining in vitro test period of different control (B), contingent (C), and noncontingent (D) preparations. E, F, Group comparisons. E, The rate of fictive biting (patterns per minute) generated in the in vitro test period remained higher in the contingent (Cont.) group than in either the control (q₂ = 4.062) or noncontingent (Non-cont.; q₃ = 3.315) groups. The control and noncontingent groups were not significantly different (q₂ = 0.878). F, The proportion of isolated preparations that generated rhythmic fictive biting and associated B63 impulse bursts was significantly higher in preparations from the in vivo contingently trained group than either the control (p < 0.05) or noncontingently trained groups (p < 0.01). Control versus noncontingent groups, not significant (n.s.).

Figure 9.

Lack of dependence of the regularization of fictive biting on cycle frequency. A–C, Samples of simultaneous extracellular and intracellular recordings of radula motor output and B63 bursting, respectively, in different in vitro preparations from control (A), contingently trained (B), and noncontingently trained (C) animals during tonic current injection (intensities indicated below each recording panel) into the B63 neuron. In control and noncontingent preparations that expressed irregular fictive biting in the absence of current injection (A, C, middle), a 2 nA depolarization of B63 (A, C, top) increased the motor pattern frequency to a level similar to that expressed spontaneously by a contingent preparation (B, top) but without inducing a regularization of motor pattern generation. Conversely, in the contingent preparation that already spontaneously expressed regularized fictive biting (B, top), a 2 nA hyperpolarization of B63 (B, middle) decreased the cycle rate to a level close to the spontaneous frequencies of control and noncontingent preparations (A, C, middle) without affecting the overall regularity of motor pattern production. For all preparations, the injection of −5 nA into B63 completely suppressed motor pattern genesis (A–C, bottom).

Figure 10.

Group comparison of the influence of B63 on the rate and regularity of radula motor pattern generation. A, Continuous depolarizing current (+2 nA) injected into the B63 neurons of control and noncontingent preparations (left and right gray bars) significantly increased the rate of fictive biting from that before current injection (left and right black bars; control, W = 26; noncontingent, W = 26). Conversely, continuous hyperpolarizing current (−2 nA) injected into the B63 neurons of contingent preparations (middle gray bar) significantly decreased the rate of motor patterns compared with that occurring spontaneously (middle black bar; W = 21). B, In the same preparations as in A, however, the proportions of control, contingent, and noncontingent preparations in which spontaneous fictive biting was rhythmically expressed (black bars) did not change significantly with B63 current injection (gray bars; χ² = 0.0), indicating that the temporal organization of the motor patterns was not dependent on cycle frequency. Cont., contingent; Non-cont., noncontingent.