The motor output system of the S1 segment of a 9‐day‐old mouse. This is a two photon image of intracellular calcium in live motoneurons. The motoneurons were loaded retrogradely with a calcium dye via the S1 ventral root. The view is from above, with the lateral margin of the cord along the bottom of the figure (see orientiation arrows). A significant portion of the ventral root is apparent in the lower left part of the image; the motoneurons lie along the lateral edge of the ventral cord. Note the range in size of cell bodies and the numerous dendritic branches. Unpublished results from Mingchen Jiang, Shuaib Ahmed, and CJ Heckman. Figure prepared, with permission, by Rochelle Bright.

Muscle unit territories in the human sartorius muscle. (A) Arrangement of recording electrodes (R1‐R5) along the length of the sartorius muscle. (B) End‐plate location (filled circles) and the extent of muscle length along which the potentials for single muscle units were recorded. Each vertical bar represents a single muscle unit. Adapted, with permission, from Harris et al. 293.

Muscle unit territories in the human sartorius muscle. (A) Arrangement of recording electrodes (R1‐R5) along the length of the sartorius muscle. (B) End‐plate location (filled circles) and the extent of muscle length along which the potentials for single muscle units were recorded. Each vertical bar represents a single muscle unit. Adapted, with permission, from Harris et al. 293.

High‐density surface electromyogram (EMG) recordings of motor unit activity in a human hand muscle. (A), surface electromyogram (EMG) signals recorded with an electrode array (5 columns × 13 rows) placed over the abductor pollicis brevis muscle of a healthy man as he sustained an isometric contraction at 10% of the maximal voluntary contraction force. The 61 electrodes were 5 mm apart and the signals were detected as bipolar recordings between adjacent electrodes. The columns were aligned with the direction of the muscle fibers. (B) Distribution of multichannel EMG amplitude at the instant indicated by the red dots and vertical lines in A. Some of the recording locations are indicated by numbers corresponding to those in A. Decomposition of the multichannel surface EMG signal identified 18 motor units, 8 of which discharged action potentials at this instant and contributed to the amplitude of the interference signal. (C) The amplitude of the action potentials for the eight motor units (MU) as detected by the array at the instant shown in B. The amplitude distribution for the interference signal (B) corresponded closely with the sum of the amplitude distributions for the eight motor units (C). The amplitude calibration (mV) refers to all amplitude maps. Adapted, with permission, from Farina et al. 217.

Voltage sensitive channels in motoneurons can generate complex effects. [(A), (B)] The dynamics of the H‐current induce a resonance in motoneuron subthreshold behavior. A shows the voltage response of the neuron (upper trace) to an injected current of steadily increasing frequency (lower trace). The voltage response grows and then declines. In B, the resulting impedance versus input frequency shows a peak at about 10 Hz. Adapted, with permission, from Manuel et al. 483. (C) L‐type Ca currents have an intrinsic tendency to generate prolonged tail currents. Increasing the duration of the voltage step increases the duration of the tail current (compare red and blue traces to black). Recordings obtained in nucleated patches of membrane from hypoglossal motoneurons. Modified, with permission, from Moritz et al. 522.

Input organization and electrical behavior of motoneurons in the base state, without significant levels of neuromodulation. (A) Summary of distribution of ionotropic synaptic currents within a motor pool. Data from a series of studies in motoneurons that innervate cat ankle extensor muscles by Powers, Binder and colleagues (reviewed in reference 579). Inputs generated by electrical stimulation of descending tracts [pyramidal (Pyr), rubrospinal (RN), vesitbulospinal (from Dieter's nucleus (RN)], ventral roots (recurrent inhibition via Renshaw cells), and tendon vibration [Ia monosynaptic excitation (Ia excit) and Ia disynaptic inhibition (Ia inhib)]. All measurements were made under steady‐state conditions. Figure modified, with permission, from reference 579). (B) Repetitive firing in response to a linearly rising injected current (control, thin trace). The thick trace indicates the response after administration of riluzole to block the NaPIC; the transient Na current remains but spiking cannot occur due to the lack of the acceleration in membrane potential due to the NaPIC (arrow on control trace). Redrawn, with permission, from the authors from data from Kuo et al. 418. (C) Summary of frequency‐current relations in S, FR, and FF motoneurons generated by linearly increasing injected currents. Based on simulations, with permission, by Heckman and Binder 305.

Distribution of neurotransmitter serotonin (5‐HT) contacts on a cat motoneuron. Each dot indicates the presence of a synaptic bouton containing 5‐HT. Reconstruction of the cell, with permission, by Dianne Dewey. Figure created from data from Alvarez et al. 15, with permission from the authors.

Effects of increasing Ca persistent inward current (CaPIC) amplitude on the firing produced by an excitatory synaptic input. A highly realistic motoneuron model was used to generate these results, based on a full reconstruction of a fast fatigue resistant (type FR) motoneuron with more than 700 dendritic compartments (based on references 194 and 193). The CaPIC was generated at a primarily middendritic location 89,193,269. On the left, a step increase in conductance is applied to excitatory synapses distributed throughout the dendritic tree. On the right, the same simulated synapses were activated in a triangular fashion. From the bottom up, the amplitude of the CaPIC is increased progressively, mimicking one of the main effects of increasing neuromodulatory input. Note the need for inhibitory input after the triangle of excitation to bring firing to a stop at the highest level. Simulations, with permission, by Dr. Sherif ElBasiouny.

Current‐voltage (I‐V) and frequency current (F‐I) relations generated by triangular input waveforms in motoneurons with a steady background of neuromodulatory input. Black arrows indicate ascending and descending phases. (A) I‐V relation in a putative type S motoneuron. The persistent inward current (PIC) onset is indicated by the red arrow. (B) F‐I relation in the same type S motoneuron. The red arrow indicates the onset of the acceleration in firing due to PIC activation, which occurs right at threshold for onset of repetitive firing. Note that the hysteresis in PIC onset and offset (blue arrow) in A generates the hysteresis in firing onset and offset in B. (C) I‐V relation in a putative FF motoneuron. (D) F‐I relation in the same FF motoneuron. The lesser degree of I‐V hysteresis corresponds to lesser F‐I hysteresis. (E) I‐V relation in a rat motoneuron. (F) F‐I relation in the same rat motoneuron. As for the cat, the I‐V hysteresis produces F‐I hysteresis. [(A)‐(D)] Modified, with permission, from Lee and Heckman 437,439. [(E), (F)] Modified, with permission, from Hamm et al. 286.

The persistent inward current (PIC) amplifies both excitatory and inhibitory synaptic inputs. (A) Amplification during repetitive firing. Lowest trace: sinusoidal muscle stretch to provide synaptic input. Next: linearly rising current injected into motoneuron to generate background firing. Upper trace: firing pattern evoked by the combined stretch and current. The strong increase in firing range in the last few cycles is due to amplification by the PIC (onset marked by green arrow). Modified, with permission, from Hultborn et al. 355. (B) Amplification of excitatory synaptic current as a function of voltage clamp holding potential. As the membrane potential is depolarized (x‐axis), synaptic current (y‐axis) increases and then decreases, demonstrating amplification and saturation. The vertical arrow indicates the approximate membrane potential at which the cell would begin spiking in unclamped conditions. The three traces show increasing levels of neuromodulatory input, low to high. Modified, with permission, from Lee and Heckman 435. (C) Computer simulation in which dendrites were depolarized by excitation alone (upper traces) and then by combined excitation and inhibition. Excitation alone activated the PIC, but inhibition sharply reduced PIC activation (red arrow). Modified, with permission, from Bui et al. 87. (D) Comparison of excitatory and inhibitory currents within one cell during voltage clamp at various membrane potentials. The excitatory current shows the same amplification pattern seen in B but inhibition deactivates the PIC and thus had its greatest influence in the voltage range where the PIC reaches its peak. Modified, with permission, from Hyngstrom et al. 358.

Contractile properties of two motor units in first dorsal interosseus as estimated with spike‐triggered averaging. (A) The subject slowly increased the abduction force exerted by the index finger during an isometric contraction and the action potentials discharged by two motor units in first dorsal interosseus were detected with an intramuscular electrode. (B) The twitch forces of the two motor units (1 and 2) were estimated with spike‐triggered averaging of 512 discharge times. The time to peak force (contraction time) was 65 ms for Unit 1 and 39 ms for Unit 2. The peak forces were 2.9 and 14.7 mN, respectively. Modified, with permission, from Desmedt and Godaux 164.

Contractile properties of 528 motor units in tibialis anterior as estimated with spike‐triggered averaging. (A) Histogram of twitch torques. (B) Relation between twitch torque and recruitment threshold for 40 motor units from a single individual. The linear relation (y = 0.72× + 13.0, r = 0.69) was statistically significant (P < 0.001). The relation for 514 motor units was slightly weaker (r = 0.61), but still statistically significant (P < 0.001). (C) Histogram of time to peak torque. (D) The relation between time to peak torque and recruitment threshold for the 40 motor units from the same subject was not statistically significant (y = –0.09 + 47.4, r = –0.23). However, there was a slight, but statistically significant, negative relation (y = –0.11× + 44.5, r = –0.21, P < 0.001) between time to peak torque and recruitment threshold for the sample of 514 motor units. Modified, with permission, from Van Cutsem et al. 723.

Relation between stimulus frequency and the force exerted by single motor units located in muscles that extend the toes. (A) Motor units activated by intraneural stimulation (upper trace) evoked a dorsiflexion force (lower trace). (B) Relation between stimulus frequency and normalized force (mean ± SE) for the 13 motor units. Modified, with permission, from Macefield et al. 474.

Spike‐triggered averaging of motor unit activity. (A) The experimental arrangement developed by Stein et al. 672 to estimate the contractile properties of motor units in first dorsal interosseus during a voluntary contraction. The isolated action potential of a motor unit was used to trigger an averaging device that accumulated a running average from the net force for a few hundred milliseconds each time an event was detected. [Modified, with permission, from McComas 492 by Joel A. Enoka.] (B) Relation (r² = 0.582) between the spike‐triggered average force and recruitment threshold force for motor units in first dorsal interosseus during a ramp contraction. Modified, with permission, from Milner‐Brown et al. 513.

Rate modulation during ramp increases and decreases in muscle force. (A) The recruitment and derecruitment thresholds of four motor units in rectus femoris, and the extent to which discharge rate was modulated (pulses per second; pps) during ramp contractions by the knee extensor muscles. The peak force during the isometric contraction was 33% maximal voluntary contraction (MVC). [Modified, with permission, from Person and Kudina 561.] (B) The instantaneous discharge rate of 42 motor units in soleus during ramp‐and‐hold isometric contractions up to MVC force. Recruitment thresholds ranged from 0.2% to 96.4% MVC force. Adapted, with permission, from Oya et al. 545.

Motor unit activity in first dorsal interosseus during brief isometric contractions at different target forces. (A) The discharge rate of 38 motor units (each line is a different motor unit) as young adults performed the isometric contractions (∼6 s) at approximately ten target forces with each motor unit. The target forces corresponded to an abduction force exerted by the index finger. The lower limit of each line indicates the minimal target forces at which the motor unit discharged action potentials repetitively. Twelve of the 38 motor units exhibited a clear plateau in discharge rate. (B) The experimental data were used to modify the Fuglevand model of motor unit recruitment and rate modulation 243 and the simulated data were compared with experimental measurements of the coefficient of variation for force across the operating range of the muscle [2‐95% maximal voluntary contraction (MVC) force]. The output of the model was indistinguishable from the experimental measurements and the graph shows the corresponding motor unit activity at each target force. Modified, with permission, from Moritz et al. 524.

Influence of contraction speed on the activity of motor units in tibialis anterior. (A) The recruitment thresholds of three motor units (indicated by different colors) when the force increased linearly to 120 N in 0.4, 1.2, 2.3, 5.0, and 10.0 s. The target force was approximately 50% maximal voluntary contraction. The recruitment thresholds were close to zero when the time to the target was less than 0.15 s. The decrease in recruitment threshold was greatest for the motor unit with the highest threshold. (B) The increase in discharge for one motor unit when the task was to reach the target in 1 s (yellow), 5 s (purple), and 10 s (green). Based on data, with permission, from Desmedt and Godaux 165.

Comparison of ballistic contractions performed with the dorsiflexor muscles before and after 12 weeks of training. (A) The torque (a) and rectified electromyogram (EMG) for tibialis anterior (B) produced by one individual during the submaximal isometric contractions. The peak torque for both traces was approximately 40% maximal voluntary contraction, but the rate of torque development was greater after training. The recruitment thresholds were close to zero when the time to the target was less than 0.15 s. The decrease in recruitment threshold was greatest for the motor unit with the highest threshold. (B) Schematic representation of the changes in dorsiflexion torque and the discharge rate of motor units in tibialis anterior (B and C). Training increased the instantaneous discharge rate (B) and the incidence of double discharges (C). Modified, with permission, from Van Cutsem et al. 722.

Discharge characteristics of human biceps brachii motor units when recruited during a sustained isometric contraction. Left, representative trains of action potentials for a motor unit in an young human when the target force was set at a small [∼5% maximal voluntary contraction (MVC); blue] and large (∼10% MVC; red) difference below the recruitment threshold force of the motor unit. Right, similar recordings for a biceps brachii motor unit in an old human. Modified, with permission, from Pascoe et al. 524.

The paired motor unit technique to estimate the contribution of persistent inward currents (PICs) to motor unit discharge. (A) Slow ramp increase and decrease in force exerted by the elbow flexor muscles. (B) The modulation of discharge rate for two motor units that were recruited during the task. The motor unit with the lower recruitment threshold, indicated as the reporter unit, provides an estimate of the synaptic drive received by the two motor units. Dashed vertical lines denote the recruitment and derecruitment thresholds of the test motor unit so that the difference in the discharge rate of the reporter unit (delta discharge frequency; ΔF) at these two instances can be determined to provide an index of the amplitude of the PIC during the ramp contractions. Modified, with permission, from Mottram et al. 526.

The range of motoneuron electrical behaviors is very wide. (A) Influence of the level of neuromodulation (low, medium, and high) on net input‐output gain of a pool of motor units. Computer simulations based on Heckman and Binder 305,307. The arrow indicates the approximate sum of synaptic currents from the excitatory synaptic input systems illustrated in Figure 5A. The simulations included increased persistent inward currents (PICs), depolarized rest potentials and decreased spike thresholds. (B) One motoneuron exhibiting two very different electrical behaviors, depending on motor task. The cell was initially depolarized during scratching, possibly activating a PIC, but then onset of strong inhibition (red arrow) eliminated the influence of the PIC. Once a new baseline was established, the oscillation of the scratch began. In the same cell, an irritating stimulus to the opposite side of the body evoked a weight support response with a sustained plateau (presumably PIC driven) with only minor oscillations. Input conductance during scratch was much greater than during weight support. Modified, with permission, from Perreault 557.

The PIC can generate saturation in firing rate. (A) In a cat extensor motoneuron, linear stretch of an extensor muscle (lower panel) produces a linear increase in synaptic current when the cell is held hyperpolarized to prevent persistent inward current (PIC) activation (thin trace, upper left panel). When the voltage clamp holding potential was depolarized to the approximate voltage at which spikes would be initiated in unclamped conditions (thick trace, upper left), the PIC amplifies (green arrow) and then saturates (red arrow) the input. (B) The firing response in the same cell to the same linear stretch, resulting in firing rate acceleration (green arrow) and rate limiting (red arrow). Data in A and B taken, with permission, from Lee et al. 446. (C) Firing pattern of a human biceps motor unit (upper panel) in response to a voluntary effort to linearly increase elbow flexion torque (bottom panel). Note the similarity to the firing pattern in the cat motoneuron in B. Data provided by Carol Mottram, with permission, from the database of reference 526.

Plasticity in motor output to chronic spinal injury. Column (A) Upper panel: testing of tail spasticity in a rat with chronic injury at the S1 level. Note the electromyogram (EMG) recording in the tail muscle (records illustrated in column B). The two images below indicate the level of neurotransmitter serotonin (5‐HT) in the rat spinal cord before and after chronic spinal injury. The “mn” in the upper image indicates the cell body of a motoneuron. The arrow in the bottom image indicates a 5‐HT containing fiber, which is rare postinjury. Column (B) tail muscle electromyograms (EMGs) recorded in response to electrical stimulation to the tip of the tail (see A, top). This brief stimulus evokes a long‐lasting reflex (LLR), which is generated mainly by motoneuron persistent inward currents. Addition of an antagonist (SB242084) to block binding of 5‐HT has no significant effect (middle EMG, blue). In contrast, an inverse agonist (SB206553), which blocks constitutive 5‐HT receptor activity, greatly reduces the LLR (red). Modified, with permission, from Murray et al. 530.

Plasticity following peripheral nerve reinnervation. (A) Illustration of the paradigm to study recovery from chronic peripheral nerve injury. Intracellular recording in a motoneuron (in ventral spinal cord, left) is used to record the responses to either electrical stimulation of the muscle nerve 1 or muscle stretch 2. (B) EPSPs generated by electrical stimulation (left column) and muscle stretch (right column). Note that some motoneurons (e.g., MN2) respond to electrical stimulation but not stretch, indicating the peripheral reinnervation of the stretch receptor in muscle has failed. Also, the response to electrical stimulation is weaker than in control (e.g., MN1). As a result, the stretch reflex does not recover. Modified, with permission, from Alvarez et al. 14.

Dramatic changes in motoneuron morphology in response to injury and disease. (A) Prolonged axotomy of adult cat motoneurons induces conversion of dendrites to axons, especially in the distal regions. The cell on the left is normal. The one on the right has undergone prolonged axotomy. The distal extensions from the dendrites, such as those into the white matter at the left, have converted to an axon‐like phenotype and exhibit significant myelination. Provided by Ken Rose from the databank of cells stained by Rose and McDermid (see also reference 469 and 470). (B) Increased dendritic branching in a mouse model of amyotrophic lateral sclerosis. A normal motoneuron is illustrated on the left, while the cell on the right is from a mutant SOD1 animal. In both cases, the animals were less than 10 days old, so these profound anatomical changes occur long before symptom onset. Different colors indicate the multiple branches originating from primary trunks extending from different parts of the soma. Modified, with permission, from Amendola and Durand 16.