Rapid Adaptation to Changing Mechanical Load by Ordered Recruitment of Identified Motor Neurons

Motor pattern and behavioral differences between unloaded and loaded swallows. Motor patterns were obtained in vivo by recording from key components of the motor system: the I2 protractor muscle, whose activation moves the grasper forward; RN, whose largest unit, motor neuron B8a/b, causes closing of the grasper following protraction (Morton and Chiel, 1993b); BN2, whose units can be identified as the motor neurons B3 (largest), B6/B9 (medium), and B38 (medium), which each activate the I1/I3 jaw muscle complex to cause retraction of the grasper (i.e., neurons B3, B6/B9; Lu et al., 2015) and to prevent food from slipping out during protraction (i.e., neuron B38; McManus et al., 2014); and BN3, whose largest unit is the multiaction neurons B4/B5, which serve an important role in the transition between protraction and retraction (Gardner, 1977; Warman and Chiel, 1995). A, Swallows are rapid in the absence of load. This record shows seven complete cycles of ingestive motor programs; with the first, the animal tried to grasp food but failed (bite); with the second, the animal successfully grasped a 5 × 0.5 cm seaweed strip and began swallowing it; by the end of the seventh cycle, the strip was fully consumed, and the animal soon after switched from swallowing to biting (unsuccessful grasping for more food). Above the neural records, “B” indicates the initial bite; bars indicate when the seaweed strip could be clearly seen in the video to be moving inward during a swallow; “S” indicates swallows in which the strip was no longer visible because it had moved completely inside the mouth. The largest unit on BN2, the B3 motor neuron, which is recruited when animals need to generate the greatest force, was activated just twice, during the third motor pattern. Note that the force record is omitted because the seaweed strip was not attached to the force transducer. B, Swallows are slowed and motor recruitment increases in response to external load. Data are obtained from the same animal as in A. At the beginning of the record, the animal was presented with a 10 × 0.25 cm strip of unbreakable tape-reinforced seaweed attached to a force transducer. With the first motor pattern, the animal bit and failed to grasp the strip; with the second, the animal succeeded in grasping and began to swallow; with the third, the strip began to grow taut, and force can be seen (dashed line indicates zero force); by the fifth, the strip was completely taut at peak retraction. The measured force then rose and fell periodically with the retraction phase of each swallow. Because the strip could not break, the animal fed near the surface of the water, with the anterior portion of its head lifting out of the water some short distance with each effort to ingest the strip; as each retraction phase ended, the animal lost its momentary progress up the strip and fell gently back to the surface of the water, with force declining during this time. As in A, video was used to track when seaweed was moving into the mouth, which is indicated with lines above the neural records; the durations of inward movement were initially much longer compared with when external load was completely absent (compare A), and consequently the overall rate of swallowing was slower. Note that although the durations of inward movement were longer, the amplitude of inward movement may not have been very different due to the large mechanical load reducing the speed of inward movement (i.e., the amplitude of inward movement is limited by the amplitude of grasper movement). In contrast to unloaded swallowing, the B3 motor neuron (largest unit on BN2) was usually very active during retraction. After the first nine swallows, the animal introduced pauses between swallowing bouts. After ∼3 min of swallowing without successfully breaking the seaweed, the animal gradually changed from tugging to a radula scraping strategy (data not shown) and eventually released and moved away from the food. For subsequent analyses, only the first, least variable swallows on unbreakable seaweed were used, excluding variation due to changing behavioral strategies, and only swallows after tension in the strip had fully developed were analyzed. C–E, Total cycle time and the duration of inward movement increase in response to load, but the time between inward movements does not. Five animals (numbered 1–5) swallowed unloaded (U) and loaded (L) seaweed strips. Cycles were partitioned by the inward movement of the seaweed strips. The total cycle time (time from start of one inward movement to the next), as well as the period of inward movement, showed statistically significant increases in response to load, whereas the time between inward movements, when the seaweed was either moving outward or was stationary, did not. Mean values are plotted for each animal; whiskers show the SEM. The asterisks above the plots indicate statistically significant increases. See Table 2 for statistical details.

Changes in burst duration and mean firing rate associated with load. A, B, Protraction phase motor neuron activity (B38 and I2) was unchanged by load. C, D, E, In contrast, the durations of retraction phase motor neuron activity (B8a/b and B3/B6/B9) and B4/B5 activity increased when load was present. F, The firing frequency of retractor muscle motor neurons (B3/B6/B9) also increased when load was present. Mean values are plotted for each animal; whiskers show the SEM. The asterisks above the plots indicate statistically significant increases in duration and firing rate. For statistical details, see Tables 3 and 4. Numbers 1-5 indicate individual animals. U, unloaded; L, loaded.

Biomechanics, timing, and frequency of motor neuronal activity during unloaded versus loaded swallowing. A schematic of the biomechanics of swallowing helps explain the changes in the timing and frequency of motor neuronal activity between unloaded and loaded swallows. A, The biomechanics and motor control of swallowing. The stages of swallowing a seaweed strip under tension are illustrated schematically in a midsagittal view of the buccal mass, with the anterior opening of the mouth at the right and the esophagus at the left. Panels correspond to the five phases observed in the force record. See Materials and Methods for force segmentation procedure, and see Results for a detailed explanation of each stage. Closing of the grasper is illustrated by a change of shape from roughly spherical (stages I, II, and V) to ellipsoidal (stages III and IV; Neustadter et al., 2002). Points of contact between the seaweed and the buccal mass are indicated by black dots. The schematic in A was modified from Cullins et al. (2015b). B1–C2, Muscle and identified neuronal activity during unloaded and loaded swallowing. B1, B2, The timing of bursts of identified motor units are plotted for swallows from five animals on unloaded seaweed strips (left; n = 4, 4, 4, 4, 7 swallows) and on anchored, unbreakable seaweed strips (right; n = 5, 15, 5, 9, 5 swallows). Boxes indicate median timing, and whiskers indicate the lower and upper quartiles for the beginnings and endings of bursts. The period of seaweed inward movement is similarly indicated. C1, C2, The firing frequencies of the units are plotted for the same datasets. Thick lines indicate median frequencies, and dashed lines indicate the lower and upper quartiles for frequency. For loaded swallows (C2), force is similarly plotted, and the drop in force at the end of the previous swallow can be seen at the start (initial stage V). To aggregate swallows, time was normalized using different methods for unloaded and loaded swallows that still permit comparison (see Materials and Methods). Vertical dotted lines indicate the boundaries of segmentation used for normalization. For unloaded swallows, behaviors were aggregated by normalizing time using the period of inward seaweed movement; inward movement timing was therefore invariant after normalization, so whiskers are omitted in the left panels. For loaded swallows, behaviors were aggregated by normalizing time using segmentation of the force record; inward movement was therefore variable after normalization, so whiskers are shown in the right panels. For unloaded swallows, the B3 retractor motor neuron was recruited so rarely that it burst in only 6 of 23 swallows; consequently its typical burst timing (B1) is not plotted, and its typical firing frequency (C1) was so low that only the upper quartile is visible. Overall, the durations of bursts of the B8a/b closure motor neurons, the combined B3/B6/B9 retractor motor neurons, and the B4/B5 multiaction neurons were longer under load; in contrast, the burst durations of protraction phase units B38 and I2 were not significantly different (Fig. 4, quantification). Likewise, the intensities of firing of the B6/B9 and B3 motor neurons during retraction were greater in the presence of load (Fig. 4, quantification).