Inhibition in synchronously firing human hippocampal neurons.

In order to study the extent of inhibition in human epileptic hippocampus, we recorded extracellular unit activities of human hippocampal neurons and their responses to single pulse stimulation in temporal lobe epilepsy patients during interictal periods. The criteria for diagnosing the hippocampus as epileptic were: (1) all seizures originated in that one hippocampus, (2) surgical removal of that hippocampus resulted in seizure relief, and (3) the surgically excised hippocampus was sclerotic. Analysis of firing pattern by cross-correlation showed that synchronized firing between neurons occurred only in the epileptic hippocampus. However, synchronized firing was not limited to only bursting neurons, as previously reported in some animal models of epilepsy, but was also observed among non-bursting neurons in the epileptic hippocampus. Furthermore, no significant difference in distribution of burst-discharge neurons was found between epileptic and non-epileptic hippocampi. In response to single pulse stimulation, neurons in both 'normal' (contralateral hippocampus) and epileptic hippocampus showed a rapid increase of firing (excitation), cessation of firing (inhibition), or a sequence of both (initial excitation followed by inhibition). However, a significant difference was found in the duration of the inhibition between synchronously firing neurons and non-synchronously firing neurons: the inhibition evoked by a single stimulation in synchronously firing epileptic neurons was significantly longer (373.8 msec +/- 35.9 S.E.M., P less than 0.005) than that of non-synchronously firing neurons (83.9 msec +/- 8.9 S.E.M.). Moreover, prolonged inhibition in synchronously firing epileptic neurons could occur with little or no prior excitation, suggesting that this inhibition does not necessarily depend on an intrinsic Ca2+-dependent K+-mediated after-burst hyperpolarization but is rather likely to be synaptic. As this inhibition was longer when epileptic neurons fired in synchrony, it could be interpreted that principal neurons recruited more recurrent inhibitory circuits by firing synchronously. By taking into account the previously reported neurophysiological evidence in human in vitro epileptic tissue showing GABA-mediated inhibition and the neuroanatomical evidence in excised human epileptic hippocampus showing GAD-positive neurons and synapses, our data suggest that, in human chronic epileptic hippocampus, recurrent inhibition remains functional, and alterations in GABA-mediated inhibition may not represent the critical change responsible for seizure generation.

Neurophysiological characterization of commissurally projecting dentate neurons in the rat.

Commissural neurons in the dentate hilus and in the deep dentate granule cell layer were recorded intracellularly in vivo, in conjunction with combined injection of the anterograde tracer Phaseolus vulgaris leucoagglutinin (PHA-L) at sites of electrical stimulation. Two hilar neurons responded with short latency antidromic spikes to stimulation of the contralateral dentate infrapyramidal molecular layer, but did not show any synaptic potentials, suggesting that these neurons do not receive commissural hilar input, either directly or indirectly, from the stimulating sites. On the other hand, 3 dentate-hilar border neurons responded to the contralateral hilar stimulation with antidromic spikes, excitatory postsynaptic potentials (EPSPs), orthodromic spikes, and inhibitory PSPs (IPSPs), suggesting a rich synaptic interaction both commissurally and locally in this region. No direct commissural inhibition was observed in any of the cells. PHA-L injection at the stimulation site indicated that commissural hilar axon terminals project to a limited region of the contralateral molecular layer in a lamellar fashion, and have only a sparse distribution in the contralateral hilus. The results indicate that rapidly conducting commissural neurons in the dentate gyrus are themselves inhibited in an indirect manner by commissural fibers.

Feedforward inhibition of the rat entorhinal cortex and subicular complex.

We used in vivo intracellular recording techniques in order to provide evidence about the source of postsynaptic inhibition in the rat entorhinal cortex and subicular complex. Several different structures in the basal forebrain and hippocampus were electrically stimulated in order to activate inhibition by different pathways. This allowed a test of 2 different neuronal circuit models: feedback inhibition, in which recurrent collaterals from principal cell axons excite a local population of inhibitory neurons, and feedforward inhibition, in which excitatory afferents activate the inhibitory neurons. In both models, inhibitory cell axons branch and contribute to the inhibition of a population of principal cells. In the feedback model, a good correlation between antidromic and inhibitory response latencies is predicted. The feedforward model predicts independent antidromic and inhibitory response latencies. In one particular model of feedforward inhibition, afferents excite both local inhibitory cells and principal cells. This model predicts a high correlation between principal cell EPSP and IPSP latencies. The results showed no consistent relationship between the presence of antidromic action potentials and the presence of inhibition in response to stimulation of different sites. In addition, there was no correlation between antidromic and inhibitory response latencies. These results provide no clear support for the feedback model of inhibition. By contrast, there was a highly significant correlation between the latency of principal cell EPSPs and IPSPs, in support of a feedforward model of inhibition. Response latencies of candidate inhibitory neurons were also consistent with the feedforward model. The results provide evidence that an excitatory relay function of the entorhinal cortex and subicular complex is modified temporally by local, extrinsically activated inhibitory circuits.

Diversity in periodic pattern of firing in human hippocampal neurons.

Firing periodicity was examined in human hippocampal neurons using autocorrelation analysis. Extracellular single-unit activities were recorded from the anterior hippocampus through fine platinum microelectrodes, and the typical firing pattern in an entire recording period was reconstructed statistically in autocorrelograms (average number of firings analyzed: 5639.0 +/- 968.1 SE, range: 1158 to 31,203; number of single-unit trains was 57). Three types of periodic firing were identified as highly consistent. The first pattern consisted of a random recurrence of high-frequency action potentials (100 to 300 Hz) and was observed as an intermittent burst. In this burst, the first 10 to 30 ms after the onset of the burst was the patterned firing of several action potentials, suggesting that the generation of this stereotyped portion of the burst is primarily due to intrinsic membrane characteristics. The second pattern was the continuous rhythmical firing with a lower frequency ranging from 1 to 30 Hz. The third pattern was a clustered rhythmical firing in which a series of short rhythmical firings recurred with regular intervals; the frequency of short rhythmical firing varied from 6.7 to 17 Hz between neurons, and the interval of the regular recurrence of these rhythmical firings ranged from 0.5 to 10 s among neurons. These firing periodicities not only cover a cellular rhythm in the theta frequency reported in the lower mammalian hippocampus but also appear to be more diverse than those previously reported for hippocampal neurons in the animal literature.

Firing patterns of human limbic neurons during stereoencephalography (SEEG) and clinical temporal lobe seizures.

Comparisons of the patterns of neuronal firing and stereoencephalography (SEEG) recorded from the same microelectrodes chronically implanted in the human limbic system were made in order to study neuronal electrogenesis at onset and during propagation of focal partial complex seizures. Alert or sleeping patients were monitored during spontaneous subclinical seizures (no alterations in consciousness detectable), during auras reported by the patients as typical, and during clinical seizures with loss of consciousness, movements and post-ictal confusion. During subclinical SEEG seizures (ipsilateral, normal consciousness), few neurons increased firing (estimated at only 7%) either at the focus or at propagated sites. During auras, with altered consciousness, there were relatively few neurons that increased firing, with the estimate about 14% or twice as many as during a subclinical seizure. During the onset of a clinical seizure that involved loss of consciousness, movements and post-ictal confusion, many neurons were recruited into increased firing, with an estimate of approximately 36%. During this increased electrogenesis, neurons fired briefly in association with high-frequency local SEEG; however, the bursts were shorter than the SEEG seizure pattern. Apparently, other local neurons were recruited to fire in bursts to sustain sufficient axonal driving for widespread propagation of the seizure. When the focal SEEG slowed, the units stopped firing, which suggested that the 'focal' seizure need not be sustained for more than several seconds because propagated seizure activity was self-sustaining at distant structures. The data lead to the conclusion that SEEG seizures can be generated focally by synchronous firing of fewer than 10% of neurons in the 'epileptic pool.' However, when greater percentages of neurons are recruited in the 'epileptic focus' there is greater propagation to widespread sites, especially contralaterally, which will produce clinical partial complex seizures.

Difference in projections to the lateral and medial facial nucleus: anatomically separate pathways for rhythmical vibrissa movement in rats.

The present paper demonstrates that the lateral and medial subdivisions of the rat facial motor nucleus (NVII) receive differing mesencephalic and metencephalic projections. In order to study brain projections to facial nucleus, horseradish peroxidase (HRP) was injected iontophoretically into the entire facial nucleus or the following subdivisions: lateral, dorsolateral, medial, intermediate, and ventral. In the mesencephalic region, the retrorubral nucleus was found to project to the contralateral medial subdivision of NVII, while the red nucleus was found to project to the contralateral lateral subdivision of NVII. Other mesencephalic projections to the facial nucleus arose from the deep mesencephalic nucleus, oculomotor nucleus, central gray including interstitial nucleus of Cajal and nucleus Darkschewitsch, superior colliculus and substantia nigra (reticular). In the mesencephalic region, the Kölliker-Fuse nucleus, parabrachial nucleus, and the ventral nucleus of the lateral lemniscus projected mainly to the ipsilateral lateral subdivision of NVII. In addition, the trapezoid, pontine reticular, vestibular, and motor trigeminal nuclei were observed to have predominantly ipsilateral connections to the facial nucleus. In contrast, projections from the myelencephalic region were to both the lateral and medial subdivision of NVII. The medullary reticular nucleus, ambiguus nucleus, spinal trigeminal nucleus and parvocellular reticular nucleus projected to both lateral and medial subdivisions of NVII with an ipsilateral predominance. The gigantocellular and paragigantocellular reticular nuclei, raphe magnus, external cuneate nucleus and the nucleus of the solitary tract also projected to the facial motor nucleus. Surprisingly, no direct projections to the NVII were observed from diencephalic and telencephalic regions. Our findings that the lateral subdivision of NVII which innervates vibrissa-pad-muscles (Dom et al. 1973; Martin and Lodge 1977; Watson et al. 1982) receives different metencephalic and mesencephalic projections than medial subdivision which controls pinna movement (Henkel and Edwards 1978), suggest that the functional difference between these subdivisions is mediated by the anatomically separate pathways. We confirmed our anatomical findings by eliciting exclusively vibrissa responses by electrical stimulation of the nuclei which project to the lateral subdivision of NVII.

Structurally stable burst and synchronized firing in human amygdala neurons: auto- and cross-correlation analyses in temporal lobe epilepsy.

Burst structure and synchronized firing of bursts were studied, in the interictal period, using auto- and cross-correlation analyses in human amygdala neurons in temporal lobe epilepsy patients diagnosed as having a unilateral limbic seizure focus in anterior hippocampus and/or amygdala. Satisfactory single unit recordings were obtained from chronically implanted microelectrodes in 51 amygdala neurons, and auto-correlation analysis identified 27 of 51 neurons where burst firings recurred with regular interspike interval structures (structurally stable burst: S-burst). This structural stability was characteristic only for a short burst, or at the beginning of a series of repetitive firings, involving 2-5 action potentials. In 'non-epileptic' amygdala neurons located contralateral to the seizure focus, the average duration of S-burst was 15 msec and the number of action potentials (spikes) in the S-burst was inversely related to the interspike intervals in the S-burst, suggesting that endogenous membrane characteristics of non-epileptic amygdala neurons determine the patterns of S-burst. In contrast, in the seizure focus amygdala ('epileptic'), the duration of the S-burst was prolonged among epileptic neurons, not because of the occurrence of more action potentials within the S-burst, but because of a prolonged interspike interval within the S-burst. Furthermore, there was no relationship between the interspike interval and the number of action potentials in the S-burst, suggesting that synaptic inputs and/or extracellular environmental factors may affect an intrinsic mechanism for generating stable S-burst in epileptic neurons. Cross-correlation analysis identified synchronized firings in epileptic neurons: when two epileptic neurons both exhibited S-bursts, when either epileptic neuron exhibited S-burst, but never when neither exhibited S-bursts. Conversely, non-epileptic neurons rarely fired synchronously; even though they showed S-bursts. The difference in the pattern of S-bursts between epileptic and non-epileptic amygdala neurons seems to be the degree of firing synchrony. Our results provide, for the first time, direct evidence that human epileptogenic amygdala neurons recorded in vivo have unique burst firing patterns and significant synchronous excitatory interactions, different from a burst pattern found in non-epileptogenic amygdala neurons during the interictal period.

Tuning the power spectrum of physiological finger tremor frequency with flickering light.

Fast Fourier Transform analyses were performed on finger tremor movements at 0.2-Hz intervals from 0.4 to 40 Hz in 10 human subjects, under a flickering light condition of 4-15 Hz and an unstimulated control condition. Under the control condition, the power spectrum showed an essentially normal curve distribution, except for an early frequency component in the histogram. In contrast, when the flickering light stimulus was presented, the power of specific frequency components at 8-11 Hz was strongly enhanced. This effect was induced exclusively at a frequency of 8, 9, or 11 Hz of flickering light, and this flickering frequency producing the enhancement effect differed from subject to subject. There existed a significant correlation between the frequencies of flicker and tremor at the tuned frequency. These findings demonstrate that a specific frequency of flickering light can intensify a specific frequency of physiological finger tremor, and that different individuals exhibit different optimal "tuning" frequencies.