Multiple Local Synaptic Modifications at Specific Sensorimotor Connections after Learning Are Associated with Behavioral Adaptations That Are Components of a Global Response Change

Abstract

Learning causes local changes in synaptic connectivity and coordinated, global changes affecting many aspects of behavior. How do local synaptic changes produce global behavioral changes? In the hermaphroditic mollusc Aplysia, after learning that food is inedible, memory is expressed as bias to reject a food and to reduce responses to that food. We now show that memory is also expressed as an increased bias to reject even a nonfood object. The increased bias to rejection is partially explained by changes in synaptic connections from primary mechanoafferents to five follower neurons with well defined roles in producing different feeding behaviors. Previously, these mechanoafferents had been shown to play a role in memory consolidation. Connectivity changes differed for each follower neuron: the probability that cells were connected changed; excitation changed to inhibition and vice versa; and connection amplitude changed. Thus, multiple neural changes at different sites underlie specific aspects of a coordinated behavioral change. Changes in the connectivity between mechanoafferents and their followers cannot account for all of the behavioral changes expressed after learning, indicating that additional synaptic sites are also changed. Access to the circuit controlling feeding can help determine the logic and cellular mechanisms by which multiple local synaptic changes produce an integrated, global change in behavior.SIGNIFICANCE STATEMENT How do local changes in synapses affect global behavior? Studies on invertebrate preparations usually examine synaptic changes at specific neural sites, producing a specific behavioral change. However, memory may be expressed by multiple behavioral changes. We report that a change in behavior after learning in Aplysia is accomplished, in part, by regulating connections between mechanoafferents and their synaptic followers. For some followers, the connection probabilities change; for others, the connection signs are reversed; in others, the connection strength is modified. Thus, learning produces changes in connectivity at multiple sites, via multiple synaptic mechanisms that are consistent with the observed behavioral change.

Keywords: Aplysia; feeding; mechanoafferents; memory; rewiring.

Figure 1.

Rejection and modulation of rejection after training. A, Rejection of a swallowed cannula. The dark band on the cannula moves away from the jaws in frames 1–8. Note that the tube translates outward. Anatomical features are indicated in frame 5. The radular surface is labeled, as are the lips and perioral zone (figure is taken from Ye et al., 2006). B, Training enhances rejection, as measured by increased rate to reject a swallowed cannula (in cm/min). Means and SEs are shown (p = 0.002, t = 3.97, df = 12, two-tailed t test; N = 7 naive animals and N = 7 trained animals), asterisk marks a significant difference.

Figure 2.

Monosynaptic connectivity of afferents to the five followers examined, and the behavioral functions of the followers. Excitation is indicated by a line, inhibition by a filled circle. A, During the protraction phase of rejection, the grasper (closed by the I4 muscle, highlighted in green) is tightly closed on the inedible material (shown schematically as a tube with a mark on it), and is protracted to push the material out of the buccal cavity using the protractor muscle (the I2 muscle, highlighted in green), pushing the anterior jaw muscles open (movements of the tube, grasper, and jaw muscles are indicated using green arrows). Closing of the grasper muscle is due to activation of the B8 motor neuron (also highlighted in green; Morton and Chiel, 1993b; Sasaki et al., 2009), and activation of the protractor muscle is due to the B61/B62 motor neurons (also highlighted in green; Hurwitz et al., 1996). B, At the onset of retraction, the B4/B5 multiaction neuron is strongly activated (indicated in green; Ye et al., 2006), inhibiting both the grasper closer motor neuron B8 (shown in red) and the jaw closer motor neuron B3 (shown in red; Lu et al., 2015). As a consequence, the open grasper can move backward through the jaws (shown in red) without their closing down on the grasper, which would otherwise cause it to close on the inedible material and pull it back into the buccal cavity (Ye et al., 2006). B31/B32 innervate the I2 muscle (Hurwitz et al., 1994, 1999), but also have a central role in deciding whether or not to initiate a feeding motor response (Hurwitz et al., 1996, 2008). Enhancements of B4/B5 activity, of B61/B62 firing, and weakening of B3 firing during retraction would bias behavior toward increased rejection.

Figure 3.

Different patterns of synaptic connectivity from S1 mechanoafferents to five followers. A, Connectivity of mechanoafferents to the 5 followers. As above, excitation and inhibition are indicated by a line and a circle, respectively. B, Distribution of PSP amplitudes in the five followers of the S1 neurons in naive animals. 0 mV (no connection), shaded white; inhibitory connections, shaded red; excitatory connections, shaded green. Note the differences in distributions of the same population of mechanoafferents to the different follower neurons. B4/B5 (N = 88 connections measured) and B31/B32 (N = 69 connections measured) do not receive any inhibitory connections. Also note the differences in the percentage of unconnected neurons, and in the different distributions of EPSP and IPSP amplitudes for the five followers (N = 102 connections measured to B3; N = 57 connections measured to B61/B62; N = 61 connections measured to B8a/b). C, Examples of monosynaptic PSPs to two follower neurons in naive Aplysia.

Figure 4.

Training biases motor activity toward rejection. A, Summary of changes in connectivity after training. As above, excitation and inhibition are indicated by a line and a circle, respectively. Followers with both inhibitory and excitatory connections are shown with both symbols. Increased excitation, green shading; increased inhibition, red shading. B–F, Mean net connectivity (average amplitude of all measured connections, both excitatory and inhibitory; unconnected neurons were also included and were given a value of zero) from S1 mechanoafferents to the five followers in ganglia from naive and trained Aplysia. SEs are shown. Significant effects are marked with an asterisk. B, Training caused a significant increase (Mann–Whitney U test, p = 0.007) in mean excitation of the connections to B4/B5 (N = 88 connections in naive animals; N = 107 in trained animals). C, Training caused a significant increase (Mann–Whitney U test, p = 0.001) in mean inhibition of the connections to B3 (N = 102 connections in naive animals; N = 104 in trained animals). D, Training caused a significant increase (Mann–Whitney U test, p = 0.001) in mean excitation of the connections to B61/B62 (N = 57 connections in naive animals; N = 49 in trained animals). E, Training caused no significant change in mean connectivity to B8a/b (N = 61 connections in naive animals; N = 53 in trained animals) after training (Mann–Whitney U test p = 0.142). F, Training caused no significant change in mean connectivity to B31/B32 (N = 69 connections in naive animals; N = 109 in trained animals) after training (Mann–Whitney U test p = 0.638).

Figure 5.

Percentage of excitatory, inhibitory, and unconnected S1 mechanoafferents to five followers. The relative changes in excitation, inhibition, or no connection in naive and trained preparations were tested with χ² tests with 1 or 2 df, based on whether naive preparations did or did not have PSPs with net inhibition to a particular follower. Note that the number of connections tested is presented in the previous figure. Significant effects are marked with an asterisk. A, Training did not cause a significant difference in the distribution of the type of connections to B4 (p = 0.78, χ² = 0.81, df = 1). Note that, after training, a single inhibitory connection was observed. This did not generate a significant change in connectivity. B, Training caused a significant difference in the distribution of the type of connections to B3 (p < 0.001, χ² = 35.10, df = 2), enhancing inhibition. C, Training caused a significant difference in the distribution of the type of connections to B61/B62 (p < 0.0001, χ² = 22.17, df = 2), enhancing excitation. D, Training caused a significant difference in the distribution of the type of connections to B8a/b (p = 0.0001, χ² = 38.14, df = 2); both the percentage of excitatory and inhibitory connections increased. E, Training did not cause a significant difference in the distribution of the connections to B31/B32 (p = 0.69, χ² = 0.73, df = 1).

Figure 6.

Mean amplitude of excitatory and inhibitory connections from S1 mechanoafferents to 5 followers in ganglia from naive and trained animals. SEs are shown. Significant effects are marked with an asterisk. A, EPSP amplitude increased in B4/B5 (p = 0.003, t = 3.64, df = 159.01) and in B61/B62 (p = 0.022, t = 2.48, df = 21; two-tailed t tests assuming unequal variance), but not in the other three followers. All 13 EPSPs to B61/B62 in naive animals were 1 mV, and therefore no error bar is shown. B, IPSP amplitude increased in B3 (p = 0.0008, t = 3.45, df = 122.359; two-tailed t tests assuming unequal variance). Note that there were no IPSPs after training in B61/B62. A single IPSP was seen in B8a/b in naive animals, whose value is not shown. In addition, no IPSPs were present in B4/B5 and in B31/B32 from naive animals, and therefore no changes in IPSP amplitude could have been observed.

Figure 7.

Changes in net connectivity after repeated spikes. All measured connections are shown (the sum of excitatory, inhibitory, and unconnected S1 neurons to each follower cell). A total of 15 stimuli were applied to the presynaptic neurons, in three bursts of five spikes each, separated by 10 s. The data show the net amplitude and SEs of the 1st, 5th, and 15th response. Note that the data for the first PSP are identical to those shown in Figure 2 (blue, naïve; purple, trained). Means and SEs are shown. A, Connectivity to B4/B5. A two-way ANOVA showed significant main effects due to training (p < 0.001, F_(1,579) = 31.616) and due to repetition (p < 0.001, F_(2,579) = 10.981), with no significant interaction (p = 0.546, F_(2,579) = 0.605). B, Connectivity to B3. A two-way ANOVA showed a significant effect due to training (p < 0.001, F_(1,612) = 78.501), with no significant effect due to repetition (p = 0.694, F_(2,612) = 0.365), and no significant interaction (p = 0.936, F_(2,612) = 0.066). C, Connectivity to B61/B62. A two-way ANOVA showed a significant increase in the overall connection due to training (p < 0.001, F_(1,312) = 16.115), as well as a significant effect due to repetition (p < 0.578, F_(2,312) = 21.824), with no significant interaction (p = 0.907, F_(2,198) = 0.098). Note that the net connection becomes inhibitory with repeated stimulation, but it is less inhibitory after training. This suggests that other sources of input, not captured in these studies, may account for some of the behavioral changes during the expression of memory. An inhibitory bias for B61/B62 would reduce the likelihood that an animal would respond to food at all, and this is consistent with other behavioral observations after animals are trained on inedible food. Rather than rejecting food, animals simply do not respond to it at all. D, Connectivity to B8a/b. A two-way ANOVA showed no significant change in connectivity due to training (p = 0.345, F_(1,528) = 0.894), but there was a significant decrease in excitation as a result of repetitive spiking (p = 0.001, F_(2,534) = 6.739), with no significant interaction (p = 0.880, F_(2,534) = 0.127). E, Connectivity to B31/B32. A two-way ANOVA showed no significant change in connectivity due to training (p = 0.503, F_(1,336) = 0. 449), and no significant effect as a result of repetitive spiking (p = 0.187, F_(2,336) = 1.687), as well as no significant interaction (p = 0.585, F_(2,336) = 0.536).