The construction of movement with behavior-specific and behavior-independent modules

Abstract

Growing evidence suggests that different forms of complex motor acts are constructed through flexible combinations of a small number of modules in interneuronal networks. It remains to be established, however, whether a module simply controls groups of muscles and functions as a computational unit for use in multiple behaviors (behavior independent) or whether a module controls multiple salient features that define one behavior and is used primarily for that behavior (behavior specific). We used the Aplysia feeding motor network to examine the two proposals by studying the functions of identifiable interneurons. We identified three types of motor programs that resemble three types of behaviors that Aplysia produce: biting, swallowing, and rejection. Two ingestive programs (biting, swallowing) are defined by two movement parameters of the feeding apparatus (the radula): one is the same in both programs (phasing of radula closure motoneurons relative to radula protraction-retraction), whereas the other parameter (protraction duration) is different in the two programs. In each program, these two parameters were specified together by an individual neuron, but the neurons in each were different (B40 for biting, B30 for swallowing). These findings support the existence of behavior-specific modules. Furthermore, neuron B51 was found to mediate a phase that can be flexibly added on to both ingestive and egestive-rejection programs, suggesting that B51 may be a behavior-independent module. The functional interpretation of the role played by these modules is supported by the patterns of synaptic connectivity that they make. Thus, both behavior-specific and behavior-independent modules are used to construct complex behaviors.

Figure 2.

Characterization of motor programs elicited by CBI-4 stimulation. A, Examples of programs elicited by CBI-4. Protraction (open bar) is monitored by activity in the I2 nerve and B31, retraction (filled bar) by depolarization of B4/5 after the termination of I2 activity, hyperpolarization in B31, and closure activity in the RN and B8. Top, Plots of instantaneous frequency of B8 during program. The motor programs in A1 and A2 were ingestive, because B8 was active primarily during retraction. The motor programs in A3 and A4 were egestive, because B8 was active primarily during protraction. A2 and A4 were similar to A1 and A3, respectively, except that there was a second B8 burst (gray bar) observed during the late part of retraction (hyper-retraction). Thus, motor programs in A2 or A4 are not considered as different types of programs from those in A1 or A3, but two different variants of the same type of program. B, Cluster analysis. Plot of average B8 frequency during protraction versus that during retraction (excluding the second B8 burst-hyper-retraction) for 175 CBI-4-elicited motor programs is shown. Three clusters enclosed by dotted lines are based on visual inspection and produced the smallest determinant of within-group covariance matrix, suggesting that the data can be separated into three clusters: ingestive, intermediate, and egestive. The distribution of data points for motor programs with or without a second B8 burst is mixed, suggesting that the occurrence of the second B8 burst is independent of B8 activity during protraction and early retraction. Resting membrane potentials (in mV): B8, -57; B31, -65; B4/5, -66.

Figure 3.

Parametric features of CBI-4-elicited motor programs. A, Comparison of CBI-4-elicited programs with or without a second B8 burst. Regardless of whether a second B8 burst was absent or present, the duration of both protraction and retraction (excluding hyper-retraction) was similar in ingestive (protraction: 3.4 ± 0.3 sec, n = 29, vs 3.4 ± 0.2 sec, n = 35, unpaired two-tailed t test, t = 0.7, p > 0.05; retraction: 4.0 ± 0.2 sec, n = 29, vs 4.7 ± 0.3 sec, n = 35, unpaired two-tailed t test, t = 0.6, p > 0.05), egestive (protraction: 10.5 ± 0.7 sec, n = 19, vs 9.8 ± 0.5 sec, n = 34, unpaired two-tailed t test, t = 1.0, p > 0.05; retraction, 7.0 ± 0.3 sec, n = 19, vs 7.1 ± 0.3 sec, n = 34, unpaired two-tailed t test, t = 0.1, p > 0.05), or intermediate (protraction: 8.4 ± 0.9 sec, n = 13, vs 7.0 ± 1.0 sec, n = 12, unpaired two-tailed t test, t = 1.2, p > 0.05; retraction: 7.5 ± 0.8 sec, n = 13, vs 7.4 ± 1.3 sec, n = 12, unpaired two-tailed t test, t = 0.02, p > 0.05) programs. In addition, regardless of whether a second B8 burst was absent or present, B8 firing frequency during protraction and retraction was similar in ingestive (protraction: 1.03 ± 0.1 Hz, n = 29, vs 1.04 ± 0.1 Hz, n = 35, unpaired two-tailed t test, t = 0.4, p > 0.05; retraction: 6.6 ± 0.3 Hz, n = 29, vs 7.4 ± 0.3 Hz, n = 35, unpaired two-tailed t test, t = 1.7, p > 0.05), egestive (protraction: 5.8 ± 0.3 Hz, n = 19, vs 5.7 ± 0.2 Hz, n = 34, unpaired two-tailed t test, t = 0.4, p > 0.05; retraction: 1.8 ± 0.15 Hz, n = 19, vs 1.7 ± 0.15 Hz, n = 34, unpaired two-tailed t test, t = 0.03, p > 0.05), or intermediate (protraction: 3.3 ± 0.3 Hz, n = 13, vs 3.4 ± 0.4 Hz, n = 12, unpaired two-tailed t test, t = 0.5, p > 0.05; retraction: 4.2 ± 0.4 Hz, n = 13, vs 4.9 ± 0.5 Hz, n = 12, unpaired two-tailed t test, t = 1.8, p > 0.05) programs. The duration of hyper-retraction was similar during ingestive (4.7 ± 0.2 sec; n = 35), intermediate (5.2 ± 0.8 sec; n = 12), or egestive (5.5 ± 0.4 sec; n = 34) motor programs (one-way ANOVA; F_(2,78) = 1.4; p > 0.05). In contrast, B8 firing frequency was higher (9.7 ± 0.5 Hz; n = 35) in ingestive programs, lower (8.1 ± 0.9 Hz; n = 12) in intermediate programs, and lowest (7.0 ± 0.4 Hz; n = 34) in egestive programs [one-way ANOVA; F_(2,78) = 7.9; p < 0.001; a Bonferroni multiple comparisons test of the three different group pairs showed significant difference between B8 firing frequency in ingestive programs and that in egestive programs (t = 4.0; p < 0.001), no significant difference between that in ingestive programs and that in intermediate programs (t = 1.6; p > 0.05), and between that in intermediate programs and that in egestive programs (t = 1.1; p > 0.05)]. B, Comparison of protraction duration in CBI-4- and CBI-2-elicited programs. Data for CBI-4-elicited programs with or without a second B8 burst were pooled. In CBI-4-elicited motor programs, the average protraction duration was shortest (3.4 ± 0.2 sec; n = 64) when programs were ingestive, longer (7.7 ± 0.5 sec; n = 25) when programs were intermediate, and longest (10.1 ± 0.4 sec; n = 53) when programs were egestive [one-way ANOVA; F_(2,144) = 124.1; p < 0.0001; a Bonferroni multiple comparisons test of the three different group pairs showed significant difference among all group pairs (p < 0.001)]. In contrast, the reverse was true for CBI-2-elicited motor programs. The average protraction duration was longest (17.9 ± 2.1 sec; n = 19) when programs were ingestive, shorter (14.3 ± 1.2 sec; n = 4) when programs were intermediate, and shortest (8.2 ± 0.8 sec; n = 19) when programs were egestive [one-way ANOVA; F_(2,39) = 37.7; p < 0.0001; a Bonferroni multiple comparisons test of the three different group pairs showed significant difference between protraction duration in ingestive programs and that in egestive programs (t = 8.6; p < 0.001) and between that in intermediate programs and that in egestive programs (t = 3.1; p < 0.05), but no significant difference between that in ingestive programs and that in intermediate programs (t = 1.9; p > 0.05)]. Ing, Ingestive; Int, intermediate; Ege, egestive. Error bars represent SEM.

Figure 1.

Multiple motor patterns that resemble feeding-related behavior in Aplysia and the networks that may underlie them. A, Classification of motor programs and their correspondence to behavior. Each cycle of a motor program, regardless of what type of motor program it is, comprises a protraction (P) phase that is followed by are traction (R) phase. Motor programs can be classified into three basic forms on the basis of phasing of activity of radula closure (C) motor neuron B8 relative to protraction-retraction. If closure activity occurs predominantly during retraction, the program is ingestive; if closure activity predominantly occurs during protraction, the program is egestive; and finally, if closure activity is moderate during both protraction and retraction, the program is intermediate. Furthermore, ingestive motor programs can be classified into two subtypes on the basis of duration of protraction. If the protraction is long, the motor program is biting-like; if the protraction is short, the program is swallowing-like. Biting-, swallowing-, and egestive-like programs likely correspond to three functionally effective forms of feeding behavior observed in intact animals. Intermediate programs have been observed when radula closure is monitored extracellularly in intact animals or intracellularly in isolated CNS (Morton and Chiel 1993a,b; Morgan et al. 2002), but the exact behavioral significance of this type of program remains to be determined. B, Schematic diagrams of the feeding circuits that produce multiple motor outputs. On the left, the circuit that is activated by CBI-2 and mediates ingestive-biting-like, egestive, and intermediate programs is depicted and is based on previous work. On the right, the circuit that is activated by CBI-4 and mediates ingestive-swallowing-like, egestive, and intermediate programs is shown for comparison and is based primarily on work described in this paper. In both circuits, motor programs are initiated when CBI-2 or CBI-4 activates B63 to generate protraction. B63 and B64 reciprocally inhibit each other and mediate the alternating activity in protraction motor neurons (PM) (e.g., B31/32) and retraction motor neurons (RM). B64 is activated and the retraction phase is generated when polysynaptic excitation (data not shown) from protraction interneurons (e.g., CBI-2 and B63) overcomes inhibitory input to B64. The type of motor programs is determined by the phasing of activity of radula closure motor neurons (CM) (e.g., B8). In rejection-like programs (CBI-2 and CBI-4), CM activity during protraction is mediated by excitation that originates in B20 and inhibition that originates in B4/5. In biting-like (CBI-2) and swallowing-like (CBI-4) programs, CM activity during retraction is mediated by fast inhibition and slow excitation that originates in B40 and B30, respectively. Furthermore, the long protraction of biting-like programs is mediated by inhibition of B64 by B40, whereas the short protraction of swallowing-like programs is mediated by slow excitation of B64 by B30. In intermediate programs (CBI-2 and CBI-4), CM activity in both protraction and retraction is mediated by coactivity of B20 and B40 or B30. Finally, B51 can be activated in some CBI-4-elicited programs to mediate hyper-retraction. Open triangle, Excitation; closed circle, inhibition; s, slow synaptic actions; zig-zagged line, electrical coupling.

Figure 4.

B30 activity enhances the ingestiveness of CBI-4-elicited programs and shortens protraction duration. A, CBI-4 stimulation induced moderate activity in B30, and the program was intermediate because B8 was active during both protraction and retraction (A1). Activation of B30 through depolarization (bar) made the CBI-4-elicited program more ingestive and shortened protraction (A2). A3, A4, Group data (n = 8) for the effect of B30 depolarization on B8 activity and protraction duration [repeated-measures ANOVA; F_(2,12) = 44.5; p < 0.0001; a Bonferroni multiple comparisons test of different group pairs showed that protraction duration during B30 depolarization was significantly different from the “before” (t = 7.5; p < 0.001) and “after” (t = 8.7; p < 0.001) groups, whereas protraction duration of the before group was not significantly different from the duration of the after group (t = 1.2; p > 0.05)]. B, In another preparation, CBI-4 stimulation induced moderate activity in B30, and the program was intermediate because radula closure in B8 (as monitored by activity in RN) was active during both protraction and retraction (B1). Hyperpolarization of B30 (bar) made the CBI-4-elicited program more egestive and lengthened protraction (B2). B3, B4, Group data (n = 4) for the effect of B30 hyperpolarization on B8 activity and protraction duration [repeated-measures ANOVA; F_(2,6) = 88.5; p < 0.0001; a Bonferroni multiple comparisons test of different group pairs showed that protraction duration during B30 hyperpolarization was significantly different from the before (t = 12.3; p < 0.001) and after (t = 10.5; p < 0.001) groups, whereas protraction duration of the before group was not significantly different from the duration of the after group (t = 1.8; p > 0.05)]. In A3 and B3, data points with small symbols represent data obtained in control conditions and those with large symbols obtained during B30 stimulation (A3) or hyperpolarization (B3). Resting membrane potentials (in mV): A: CBI-4, -64; B30, -57; B63, -58; B8, -65; B: CBI-4, -61; B30, -55; B4/5, -69. Error bars represent SEM.

Figure 5.

B30 produces fast IPSPs and a slow EPSP in B8. A, Stimulation of B30 with brief current pulses at 10 Hz for different durations elicited fast IPSPs and a slow EPSP. The slow EPSP was absent (A1) or small (A2) when B30 was stimulated for brief periods but became obvious when B30 was stimulated for 3 sec (A3). Recordings were made in high-divalent saline. B, Stimulation of B30 at 20 Hz for 5 sec increased the B8 firing frequency when the current pulse in B8 (bar) was injected after B30 activity (from left to right: 12, 16, and 12 spikes in B8). For group data (n = 4), the number of spikes in B8 was as follows: before B30, 13.3 ± 1.1; with B30, 18.3 ± 1.1; after B30, 13.5 ± 1.0 [repeated-measures ANOVA; F_(2,6) = 32.7; p < 0.001; a Bonferroni multiple comparisons test of different group pairs showed that the number of B8 spikes with B30 activation was significantly different from that in the before (t = 7.2; p < 0.01) and the after (t = 6.8; p < 0.01) groups, whereas the number of B8 spikes in the before group was not significantly different from that in the after group (t = 0.4; p > 0.05)]. Recordings were made in normal saline. Voltages to the right of the records of individual neurons indicate membrane potentials.

Figure 6.

B30 elicited slow EPSPs in retraction-interneuron B64. A, Monosynaptic connections from B30 to B64. A1, The slow EPSPs in B64 elicited by B30 were more prominent when B64 was more depolarized. At a low firing frequency of B30, only small, fast IPSPs that followed B30 spikes one-for-one were obvious (A2). The slow EPSPs became obvious when B30 was stimulated with a train of spikes at a faster rate (A1). Recordings were made in high-divalent saline. B, B30 activity enhanced B64 excitability and plateau generation. Without B30 activity (B1, B4), a 3 sec subthreshold current pulse in B64 (bar) did not induce spiking in B64. When B30 was stimulated at 20 Hz for 3 sec immediately preceding the current pulse in B64, B64 fired a burst of spikes and generated a plateau potential that outlasted the current pulse injected in B64 (B2). When B30 was stimulated at 20 Hz, 3 sec before and throughout the current pulse in B64, B64 generated a plateau potential with a brief delay, and the plateau outlasted the current pulse injected in B64 (B3). Recordings were made in normal saline. Voltages to the left or right of the records of individual neurons indicate membrane potentials.

Figure 7.

Synaptic connectivity of CBI-4 and B30. A, Stimulation of CBI-4 elicited fast monosynaptic EPSPs in B30, B63, and B20 that followed presynaptic spikes one-for-one. Recording was made in high-divalent saline. B, B30 may be electrically coupled to B63 because current injection in B63 induced hyperpolarizing response in B30 (normal saline). B64 elicited fast monosynaptic IPSPs in B30 that followed presynaptic spikes one-for-one (high-divalent saline). Voltages to the left of the records of individual neurons indicate membrane potentials.

Figure 8.

B51 is both necessary and sufficient for the second B8 burst in CBI-4-elicited motor programs. A, CBI-4 stimulation elicited an ingestive program with a second B8 burst (gray bar) (A1). B51 hyperpolarization (bar) prevented the occurrence of the second B8 burst (A2). Importantly, in these ingestive programs (n = 3), B8 firing frequency during protraction (1.1 ± 0.2 vs 0.8 ± 0.1 Hz; paired two-tailed t test; t = 1.1; p > 0.05) or during early retraction (9.4 ± 0.2 vs 9.3 ± 0.2 Hz; paired two-tailed t test; t = 0.2; p > 0.05) was not influenced by B51 hyperpolarization. B, CBI-4 stimulation elicited an egestive program with a second B8 burst (gray bar; as monitored by activity in RN) (B1). B51 hyperpolarization (bar) prevented the second B8 burst from occurring (B2). Similarly, in these egestive programs (n = 4), B8 firing frequency during protraction (5.5 ± 0.7 vs 5.7 ± 0.8 Hz; paired two-tailed t test; t = 1.6; p > 0.05) or during early retraction (1.4 ± 0.4 vs 1.2 ± 0.5 Hz; paired two-tailed t test; t = 0.9; p > 0.05) was not influenced by B51 hyperpolarization. C, An egestive motor program without a second B8 burst was induced by CBI-4 stimulation (C1). Injection of a depolarizing current into B51 (bar) induced B51 firing and a second burst of B8 firing (gray bar) at the end of retraction (C2). Similarly, in these programs (n = 3), B8 firing frequency during protraction (6.3 ± 0.6 vs 6.1 ± 0.5 Hz; paired two-tailed t test; t = 0.9; p > 0.05) or during early retraction (1.2 ± 0.3 vs 1.4 ± 0.3 Hz; paired two-tailed t test; t = 1.4; p > 0.05) was not influenced by B51 depolarization. Resting membrane potentials (in mV): A: CBI-4, -63; B51, -57; c-B51, -63; B8, -59; B: CBI-4, -60; B64, -74; B51, -66; c-B51, -61; C: CBI-4, -60; B51, -64; c-B51, -65; B8, -62.

Figure 9.

B51 induces activity resembling the second B8 burst in CBI-2-elicited motor programs. A, CBI-2 stimulation elicited an ingestive program without a second B8 burst (as monitored by activity in RN) (A1), as is normally the case for CBI-2-elicited programs. Current injection into B51 (bar) induced B51 firing and a second B8 burst (gray bar) at the end of retraction (A2). Notice that B64 was active throughout retraction and hyper-retraction (the second B8 burst). B, Group data (n = 5) of retraction duration during CBI-2-elicited ingestive programs in episodes before, during, and after B51 stimulation. In cycles during which B51 was stimulated, we measured the duration of B8 activity in the absence of B51 activity (black bar; presumably the first B8 burst, 4.6 ± 0.9 sec) and B8 activity in the presence of B51 activity (gray bar; presumably the second B8 burst, 5.9 ± 0.9 sec). We performed statistical analyses on the data and found that when the durations of B8 activity in before (4.8 ± 0.8 sec) and after (4.9 ± 0.7 sec) groups was compared with the duration of the first B8 burst in episodes with B51 stimulation, there was no statistically significant difference among these group data (repeated-measures ANOVA; F_(2,8) = 0.4; p > 0.05). In contrast, when durations of the first and second B8 burst in cycles with B51 stimulation were added together (10.6 ± 1.1 sec) and then compared with those in the before and after groups, there was a statistically significant difference among these group data [repeated-measures ANOVA; F_(2,8) = 31.3; p < 0.001; a Bonferroni multiple comparisons test of different group pairs showed that the total retraction duration was significantly different from that in the before (t = 6.9; p < 0.001) and after (t = 6.8; p < 0.001) groups, whereas the retraction duration in the before group was not significantly different from that in the after group (t = 0.1; p > 0.05)], suggesting that B51 indeed caused activation of the additional B8 activity that can be considered as the second B8 burst. Resting membrane potentials (in mV): CBI-2, -62; B51, -68; B64, -73.

Figure 10.

Circuit diagrams illustrating the construction of distinct motor programs through combinations of behavior-specific and behavior-independent modules. For clarity, we show different functional configurations that include only neurons with activity that predominates (for details, see Fig. 1). A, The initiation of motor programs occurs when the protraction interneuron B63 is activated by the higher order interneurons, CBI-2 (Hurwitz et al., 2003) or CBI-4. B63 and the retraction-interneuron B64 reciprocally inhibit each other and mediate the alternating protraction (P)-retraction (R) cycles. Because they are active in all motor programs, both B63 and B64 are behavior-independent modules. B, The critical parameters for biting-, swallowing-, and rejection-like programs are encoded by three separate behavior-specific modules: B40, B30, and B20. B40 appears to be responsible for B8 (C) activity phasing during retraction and long protraction duration for biting. B30 appears to be responsible for B8 (C) activity phasing during retraction and short protraction duration for swallowing. B20 appears to be responsible for B8 (C) activity phasing during protraction for rejection. C, A behavior-independent module, B51, encodes the second B8 burst (hyper-retraction) (H-R) that can be flexibly added to a subset of swallowing- and rejection-like programs elicited by CBI-4. The interneurons and the parameters that they encode are color-coded. Open triangle, Excitation; closed circle, inhibition; s, slow synaptic actions; zig-zagged line, electrical coupling.