Cells of origin of long descending propriospinal fibers connecting the spinal enlargements in cat and monkey determined by horseradish peroxidase and electrophysiological techniques.

The cells of origin of the long descending propriospinal tract (LDPT) in the cervical enlargement were studied in cat and monkey by using the retrograde transport of horseradish peroxidase (HRP). Their distribution was confirmed electrophysiologically in cat by recording their antidromic action potentials. In cats and monkeys unilateral injections of HRP were made into the gray matter of the lumbosacral enlargement, but there was some spread to the contralateral side. In cats labeled somas were found in greatest numbers in lamina VIII and medial lamina VII, bilaterally. Labeled cells also were found bilaterally in laminae I, IV--VI, and X, but few were in IV and VI. Those in lamina V were usually in the lateral part of the lamina near the reticulated region. The cross-sectional areas of 20 neurons from each of laminae I and V--VIII were measured. Cells in lamina I were smallest and the largest were in VII and VIII. In cats with the spinal cord hemisected between the injection site and the cervical enlargement containing the somas, the bilaterality of the LDPT neurons in laminae VII and VIII was confirmed anatomically and physiologically. Contralaterally projecting neurons in laminae VIII and medial VII constituted a majority of LDPT cells in those laminae. The LDPT neurons in the dorsal horn appeared to project mainly ipsilaterally, but the number of labeled dorsal horn cells in these preparations was small. The distribution of antidromically localized cells of the LDPT was found to be in good agreement with the anatomical results. Their conduction velocity was 59 +/- 22 m/s (mean +/- s.d., n = 245). Histograms of the conduction velocity by laminae are given. In monkey the distribution of labeled somas was similar to that in the cat, except that the concentration of labeled somas in the ventral horn was more medially and dorsally located. Labeled somas were found bilaterally in laminae I, IV--VIII, and X, but more appeared to be ipsilateral to the side of the injection, especially in the dorsal horn. The bilaterality of the LDPT in the monkey was not tested with hemisections of the spinal cord. Neurons of the LDPT are ideally situated for conveying sensory information from the forelimb for eliciting reflexes in the hindlimb, as has been observed after stimulating afferents in the forelimb, and for coordinating, in general, motor functions between the two pairs of limbs.

Responses of long descending propriospinal neurons to natural and electrical types of stimuli in cat.

Long descending propriospinal (LDP) neurons (antidromically identified) having cell bodies of origin in the cervical enlargement and projecting axons at least as far as the L2 segment were studied. Extracellular recording of responses to natural and electrical stimuli was done in high-spinal cats. (1) A receptive field for natural stimuli was found for 123 LDP neurons. An additional 108 LDP cells were not activated by the natural stimuli used, but some of these fired spike potentials in response to electrical stimulation of peripheral nerves of the forelimb. There was no distinction between neurons activated and those not activated by natural stimuli on the basis of location or conduction velocity. (2) The most effective natural stimuli were mechanical manipulation of the skin (both low and high threshold), movement of joints of the paw, and pressure to the deep tissues, especially to the extensor side of the arm. These modalities of stimuli were most often excitatory, but could be inhibitory as well. (3) On the basis of modality, 4 subgroups of LDP cells were identified: those which were responsive only to mechanical-cutaneous, joint-movement, or deep-pressure stimuli, and those which responded to several of these modalities of stimuli, the multimodal group. These subgroups could not be distinguished on the basis of conduction velocity. (4) The receptive fields varied in size from small (one digit) to large (all of a forelimb). For single LDP cells they included ones with single and/or multimodal input from one or both forelimbs and various combinations of excitation and/or inhibition. However, those in the dorsal horn had only ipsilateral receptive fields, mainly of the mechanical-cutaneous type. Cells with bilateral receptive fields were mainly located medially in the ventral gray in laminae VII and VIII. (5) A comparison of the location of the subtypes of LDP cells revealed that neurons activated by mechanical-cutaneous stimuli were in laminae I and IV-VIII; whereas deep-pressure and multimodal activated neurons were almost exclusively in laminae VII and VIII. (6) LDP cells receiving input from deep-pressure receptors of the paw probably relay position or weight-bearing information about the forelimbs to the lumbosacral spinal cord. This arrangement suggests that LDP neruons function in interlimb coordination and would be active during locomotion. They probably participate also in other reflexes elicited by cutaneous and deep stimuli.

A computerized discriminator for action potentials.

A system for recognizing extracellular action potentials in the presence of noise and other pulses was developed for a PDP11/23 computer. Data can be processed either on-line or off-line from a tape recorder. There are 4 parts: (1) An RC bandpass filter attenuates noise. (2) The data input program digitizes the signal every 55 mu sec. If the signal exceeds a threshold, 12 samples of the signal and the time are written onto the disk. Up to 300 pulses/sec can be processed. (3) The pulse discriminating program recognizes an action potential by fitting the 12 points with this function: v(t) = (a + bt + ct2) exp(-t/tau). It is biphasic, has two zeroes, and decays to zero at long times. There are 4 parameters, a, b, c and tau. The operator chooses a value of tau giving good fits for all pulses of this neuron. For each pulse the computer determines values of a, b and c by the least squares technique. The zeroes of v(t) are computed. For acceptance, the total pulse height (difference between extreme voltages), the position of the first zero, and the time between zeroes must be within limits. The 12 points and v(t) can be displayed. Summary statistics and histograms are printed. It takes 80 msec/pulse for these off-line computations. (4) Occasionally a pulse may be missed or an extra one recorded. The pulse insertion-deletion program edits the disk files. The instantaneous-interspike-frequency plot is displayed and the operator moves cursors to insert or delete pulses.

A 200-point analog computer of average transients for electrophysiological measurements.

A 200-point computer of average transients uses capacitors as the memory devices and MOS-FET switches. The averaged wave form is stored exponentially and the memory decays exponentially when that input is removed. The memory exponential time constant (i.e. the number of sweeps averaged) can be varied with one resistor. Read-out to an oscilloscope occurs simultaneously with read-in of each new signal. This linear circuit computes a highly accurate average, even though 10% components are used. It is more cost-effective than commercial digital averagers for many applications and can average evoked potentials, field potentials, weak single units and intracellularly recorded postsynaptic potentials.

Pontomedullary reticular projections into the region of the ascending medial longitudinal fasciculus in cat.

The medial longitudinal fasciculus (MLF) and immediately adjacent reticular formation were stimulated electrically just caudal to the trochlear nucleus in the cat. In the pontomedullary reticular formation within 1.5 mm of the midline, antidromically excited neurons were detected in (1) the contralateral abducens nucleus, (2) the contralateral reticular formation just caudal and caudoventral to the abducens nucleus, and (3) the ipsilateral nucleus reticularis pontis caudalis beneath and rostral to the abducens nucleus. The possible involvement of these neurons in eye or body movements is discussed.

Monosynaptic activation of long descending propriospinal neurons from the lateral vestibular nucleus and the medial longitudinal fasciculus.

In decerebrate cats long descending propriospinal (LDP) neurons were recorded extracellularly in the cervical enlargement. They were identified antidromically by spinal cord stimulation at the L1-L2 level. Inputs to these cells were tested by stimulating the medial longitudinal fasciculus (MLF) 5 to 6 mm rostral to the obex, the lateral vestibular nucleus (LVN), the upper MLF 1 mm caudal to the trochlear nucleus, and the medial vestibular nucleus (MVN), all on the ipsilateral side. Action potentials were elicited in 44% (64/144) of LDP neurons in the ventral horn (laminae VII, VIII) at a segmental latency of 1 ms or less following brain stem stimulation. This was considered to be a monosynaptic latency. The most effective stimulation sites were the MLF and the LVN. MLF stimulation accounted for about two-thirds of the monosynaptically elicited action potentials and LVN for about one-third. Another 22% of LDP neurons responded at longer latencies, but some of those responses may also have been monosynaptic. Stimulation of the upper MLF and the MVN were much less effective, indicating that the MLF input was predominantly from fibers originating in the medullary and/or pontine reticular formation.