Peak discrimination as a method for quantitative evaluation of neural activity by computer

A computer method (PDP-12) for the quantification of neural activity in single- as well as few-fibre preparations of nerves was devised and tested on function-generator signals and on neurograms from baro- and chemoreceptor fibres of the carotid sinus nerve of rabbits. The principle of the program was to search continuously for maxima and minima of the electrogram regardless of the absolute voltage range in which they occurred. Action potentials were then distinguished from noise by comparing these maxima with amplitude discriminators. The latter were derived from the previous analysis of noise in selected neurograms and in recordings during activity-free intervals of the normal electrogram or temporary cold block of the preparation. From peaks accepted as action potentials the program computed total mean frequency, amplitude histogram and interval histogram as standard outputs. The method permitted one to determine in few-fibre preparations not only total activity but also response characteristics of different fibre components. However, owing to principal limitations shown to be inherent in such techniques, the exact quantification of neurograms from several fibres became impossible if the mean frequency—depending on the type of activity—exceeded several hundred to about 1000 impulses/s. By systematic variation of program parameters and type and rate of neural activity, the performance characteristics of the program as well as certain general properties of single-, few- and multi-fibre neurograms were evaluated.     

Modeling neural oscillations

A brief review of oscillatory activity in neurons and networks is given. Conditions required for neural oscillations are provided. Three mathematical methods for studying the coupling between neural oscillators are described: (i) weak coupling, (ii) firing time maps, and (iii) leaky integrate-and-fire methods. Several applications from macroscopic motor behavior to slice phenomena are provided.     

Failure of synthetic polynucleotides to affect the immunogenicity of mycobacterial ribonucleic Acid and ribosomal protein preparations

The synthetic polyribonucleotides adenylic acid (poly A), uridylic acid (poly U), cytidylic acid (poly C), and inosinic acid (poly I), whether single- or double-stranded (poly A:U, poly I:C), cannot replace mycobacterial ribonucleic acid (RNA) in the production of a high immune response in CF-1 mice against tuberculous disease. These conclusions are based on the results of several types of experiments. (i) Poly A and poly U, used either singly or in combination, did not increase the immunogenicity of mycobacterial RNA preparations whether emulsified in Freund's incomplete adjuvant (FIA) or not emulsified. (ii) Mycobacterial ribosomal protein, extracted with 2-chloroethanol, was not immunogenic; the addition of poly A:U to the protein did not produce an immune response and FIA did not affect these results. (iii) The RNA left after the protein was extracted was partially immunogenic when emulsified in FIA even though it was partially degraded. (iv) Mycobacterial RNA prepared with ethyl alcohol and partially degraded with ribonuclease had a significantly lower immunogenic activity, and the original higher immune response was not restored by the addition of poly A:U. (v) Mycobacterial RNA totally degraded by weak alkali was not immunogenic, the original immunogenic activity was not restored by the addition of poly A:U or poly I:C, and FIA again did not influence the results. These findings suggest that (i) protein, polypeptides, or other antigenic fragments, if present, are not the specific immunogens; and (ii) mycobacterial RNA is responsible for the high immunogenic activity of mycobacterial ribosomal and RNA preparations. In addition, since the double-stranded forms of these synthetic polynucleotides markedly potentiate the formation of circulating antibodies, these results also reemphasize the lack of correlation between conventional antibody formation and immunity against tuberculosis.     

Neural correlates of task and source switching: Similar or different?

Controlling everyday behaviour relies on the ability to configure appropriate task sets and guide attention towards information relevant to the current context and goals. Here, we ask whether these two aspects of cognitive control have different neural bases. Electrical brain activity was recorded while sixteen adults performed two discrimination tasks. The tasks were performed on either a visual input (letter on the screen) or self-generated information (letter generated internally by continuing the alphabetical sequence). In different blocks, volunteers either switched between (i) the two tasks, (ii) the two sources of information, or (iii) tasks and source of information. Event-related potentials differed significantly between switch and no-switch trials from an early point in time, encompassing at least three distinct effects. Crucially, although these effects showed quantitative differences across switch types, no qualitative differences were observed. Thus, at least under the current circumstances, switching between different tasks and between perceptually derived or self-generated sources of information rely on similar neural correlates until at least 900 ms after the onset of a switch event.     

The basic mechanism for the electrical stimulation of the nervous system

Neural signals can be generated or blocked by extracellular electrodes or magnetic coils. New results about artificial excitation are based on a compartmental model of a target neuron and its equivalent electrical network, as well as on the theory of the generalized activating function. The analysis shows that: (i) in most cases, the origin of artificial excitation is within the axon and the soma is much more difficult to excite; (ii) within the central nervous system, positive and negative threshold currents essentially depend on the position and orientation of the neurons relative to the applied electric field; (iii) in several cases, stimulation with positive currents is easier; and (iv) it should be possible to excite synaptic activity without the generation of propagating action potentials. Furthermore, the theory of the generalized activating function gives hints to understanding the blockage of neural activity.     

Cortical motor potential alterations in chronic fatigue syndrome.

Premovement, sensory, and cognitive brain potentials were recorded from patients with Chronic Fatigue Syndrome (CFS) in four tasks: i) target detection, ii) short-term memory, iii) self-paced movement, and iv) expectancy and reaction time (CNV). Accuracy and reaction times (RTs) were recorded for tasks i, ii, and iv. Results from CFS patients were compared to a group of healthy normals. Reaction times were slower for CFS patients in target detection and significantly slower in the short-term memory task compared to normals. In target detection, the amplitude of a premovement readiness potential beginning several hundred milliseconds prior to stimulus onset was reduced in CFS, whereas the poststimulus sensory (N100) and cognitive brain potentials (P300) did not differ in amplitude or latency. In the memory task, a negative potential related to memory load was smaller in CFS than normals. The potentials to self-paced movements and to expectancy and RT (CNV) were not different between groups. The findings in CFS of slowed RTs and reduced premovement-related potentials suggest that central motor mechanisms accompanying motor response preparation were impaired in CFS for some tasks. In contrast, measures of neural processes related to both sensory encoding (N100) and to stimulus classification (P300) were normal in CFS.     

Serotonin-immunoreactive varicosities in the cell body region and neural sheath of the snail, Helix pomatia, ganglia: an electron microscopic immunocytochemical study

The distribution and connections of serotonin-immunoreactive fibers in the cell body region and neural sheath of the central ganglia of the snail, Helix pomatia, have been examined. The cell body region of the ganglia is supplied by an extremely dense network of varicose serotonin-immunoreactive fibers which surround neuronal perikarya in the ganglia. Immunoreactive processes also run to the neural sheath of both the ganglia and the peripheral nerve roots, forming a dense network. Electron microscopy revealed five different connections of serotonin-immunoreactive varicosities, according to their target: (i) non-specialized contacts with neuronal perikarya; (ii) non-specialized contacts with axon processes on the surface of the peripheral nerve roots; (iii) non-specialized neuromuscular connections with smooth muscle fibers in the neural sheath; (iv) varicosities engulfed by glial processes in both the cell body region and neural sheath; (v) varicosities embedded in the connective tissue elements of the sheath either partly or completely free of glial processes. In all cases of appositions no membrane specializations could be observed on either site of the contacts. These observations provide morphological evidence for non-synaptic regulatory actions of serotonin-containing neurons in Helix central nervous system: (i) modulation of the activity of neuronal perikarya; (ii) involvement in neuromuscular regulation; (iii) neurohormonal modulation of peripheral processes by release through the neural sheath.     

Spike generation from dorsal roots and cutaneous afferents by hypoxia or hypercapnia in the rat in vivo

The present study aimed at investigating the responsiveness of different parts of the primary afferent neurones to a brief hypoxia, hypercapnia or ischaemia under in vivo conditions. Action potentials were recorded in separate groups of anaesthetized rats from (i) the peripheral end of the central stump of the cut L3, L4 or L5 dorsal root (dorsal root preparation); (ii) the central end of the peripheral stump of the cut saphenous nerve (saphenous-receptor preparation); (iii) the distal end of a segment of the saphenous nerve cut at both ends (axon preparation). In paralysed animals interruption of artificial ventilation for 20-60 s elicited or increased the frequency of action potentials in both the dorsal root and saphenous-receptor preparations. Activation of these preparations was also achieved by inspiration of gas mixtures containing 10-0% oxygen (mixed with nitrogen) or 20-50% carbon dioxide (mixed with oxygen) which elicited in the blood a decrease in PO2 or an increase in PCO2 with a fall in pH. Occlusion of the femoral artery for 3 min also caused spike generation in the saphenous-receptor preparations with little alteration in blood pressure. All these stimuli failed to evoke action potentials in the axon preparations. Systemic (300 mg kg-1 s.c.) or perineural (2%) capsaicin pretreatment failed to inhibit the effect of hypoxia, hypercapnia or ischaemia, indicating a significant contribution of capsaicin-insensitive neurones to the responses. It is concluded that central and peripheral terminals but not axons of primary afferent neurones are excited by a brief hypoxia or hypercapnia and the peripheral terminals by a short local ischaemia as well. Excitation of central terminals by hypoxia or hypercapnia revealed in this way an antidromic activation of dorsal roots in response to natural chemical stimuli.     

Murine Typhus Toxin: Studies on Identification of the Neutralizing Factor Present in Normal Human Serum

Studies were made on the isolation and identification of the Rickettsia typhi toxin-neutralizing factor (TNF) previously demonstrated in normal human serum. By means of various methods of separating serum proteins, such as filtration on Sephadex G-200, dextran precipitation, hydroxyapatite chromatography, and ultracentrifugation, TNF was found to be closely associated with purified serum beta-lipoprotein, although no serological relationship with this protein was demonstrated. Lipase as well as trypsin digestion of purified preparations of beta-lipoprotein destroyed the TN activity. No evidence was obtained for an association of TNF with the immunoglobulins or with any serum protein other than beta-lipoprotein. Further studies revealed that (i) serum specimens with TN titers of 1:1024 and others with titers of 1:8 or less contained the same concentration of beta-lipoprotein; (ii) purified preparations of beta-lipoprotein isolated from TNF positive and negative sera, and which had the same protein concentration, differed as much as 250-fold in TN titer; and (iii) the TN activity of a serum could be removed by absorption with antiserum to beta-lipoprotein from a positive donor, but not with antiserum to beta-lipoprotein from a negative donor.     

The median preoptic nucleus: front and centre for the regulation of body fluid,sodium, temperature,sleep and cardiovascular homeostasis

Located in the midline anterior wall of the third cerebral ventricle (i.e. the lamina terminalis), the median preoptic nucleus (MnPO) receives a unique set of afferent neural inputs from fore‐, mid‐ and hindbrain. These afferent connections enable it to receive neural signals related to several important aspects of homeostasis. Included in these afferent projections are (i) neural inputs from two adjacent circumventricular organs, the subfornical organ and organum vasculosum laminae terminalis, that respond to hypertonicity, circulating angiotensin II or other humoural factors, (ii) signals from cutaneous warm and cold receptors that are relayed to MnPO, respectively, via different subnuclei in the lateral parabrachial nucleus and (iii) input from the medulla associated with baroreceptor and vagal afferents. These afferent signals reach appropriate neurones within the MnPO that enable relevant neural outputs, both excitatory and inhibitory, to be activated or inhibited. The efferent neural pathways that proceed from the MnPO terminate on (i) neuroendocrine cells in the hypothalamic supraoptic and paraventricular nuclei to regulate vasopressin release, while polysynaptic pathways from MnPO to cortical sites may drive thirst and water intake, (ii) thermoregulatory pathways to the dorsomedial hypothalamic nucleus and medullary raphé to regulate shivering, brown adipose tissue and skin vasoconstriction, (iii) parvocellular neurones in the hypothalamic paraventricular nucleus that drive autonomic pathways influencing cardiovascular function. As well, (iv) other efferent pathways from the MnPO to sites in the ventrolateral pre‐optic nucleus, perifornical region of the lateral hypothalamic area and midbrain influence sleep mechanisms.     

Slow inhibitory potentials in the teleost Mauthner cell

In vivo recordings from Mauthner cells in adult zebrafish (Danio rerio) and goldfish (Carassius auratus) preparations with potassium chloride filled electrodes revealed a new class of long-lasting synaptic events in these cells. Their decay time constant ranged from 20 to 80ms, which is about 20 times longer than that of previously identified fast glycinergic inhibitory postsynaptic potentials in this neuron. The average time to peak of these slow events ranged from 1 to 6ms. We demonstrated that they are also inhibitory since (i) they were resistant to antagonists of the excitatory glutamatergic receptors; (ii) their amplitude was increased following chloride loading of the Mauthner cell; (iii) their reversal potential was the same as that of fast, glycinergic inhibitory postsynaptic potentials; and (iv) they produced an inhibitory shunt of the cell's membrane resistance. Furthermore, as with the fast inhibitory postsynaptic potentials, the decay time of the slow events is voltage dependent, increasing when the Mauthner cell is depolarized. However, these inhibitory postsynaptic potentials had a different pharmacological profile to the fast glycinergic ones. That is, they persisted in the presence of strychnine at doses that abolished the fast ones and they were more sensitive to bicuculline. These data are compatible with the notion that these inhibitory postsynaptic potentials are mediated by activation of a different inhibitory receptor type, and may be GABAergic. In addition, the decay time constant of the fast inhibitory postsynaptic current was shorter than the first of the two components that contribute to the bi-exponential decay reported previously for miniature inhibitory postsynaptic currents in Mauthner cells of larval zebrafish. This suggests developmental modifications and/or a switch in the assembly of glycine receptor subtypes. While amplitude distributions of the fast miniature inhibitory postsynaptic potentials recorded in the presence of tetrodotoxin generally could fit with a single Gaussian function, the amplitude histograms of slow miniature events were skewed, often with multiple nearly equally spaced peaks, consistent with the synchronous release of several quantal units. These previously undescribed slow unitary inhibitory postsynaptic potentials contribute to inhibitory synaptic noise recorded in the Mauthner cells. Specifically, autocorrelation analysis revealed gamma-like rhythms (30-80Hz) in each of two phases, characterized as "noisy" and "quiet", and dominated by the fast and slow inhibitory postsynaptic potentials, respectively. The major frequencies of these two states were significantly different (i.e. around 90 and 40Hz, respectively), suggesting that the fast and slow inhibitory postsynaptic potentials are derived from different inhibitory networks. Chloride-filled Mauthner cells gradually hyperpolarized in the presence of tetrodotoxin, reflecting the effect of ongoing activity in the interneurons that produce the slow events. We conclude that this new class of inhibitory postsynaptic potentials contributes to the tonic inhibition which controls the Mauthner cell's excitability. In physiological conditions, this regulatory influence is expressed as a continuous shunt of this neuron's input resistance and responsiveness to sensory inputs.     

Evolution''s ‘missing link’: a hypothesis upon neural plasticity, prefrontal working memory and the origins of modern cognition

Many activities such as reading, mathematics and chess depend upon cognitive processes which arose after our evolution. Why could they arise if not evolved? I argue four things fortuitously came together to make our nonevolved cognitive skills possible: (i) neural plasticity; (ii) large functionally uncommitted prefrontal, temporal and parietal cerebral cortices; (iii) the ability of their neural circuits (due to neural plasticity), if trained, to take on novel symbolic and nonsymbolic skills; and (iv) a large prefrontal cortex which could use its working memory as a tuition management sketch pad in which to train them. Pre-evolved for other reasons, these four (together with invented symbolic systems and technology) together enable modern humans to ‘upgrade’ our already evolved cognitive skills to do new and nonevolved things.     

Review and current status of SPECT scatter correction

Detection of scattered gamma quanta degrades image contrast and quantitative accuracy of single-photon emission computed tomography (SPECT) imaging. This paper reviews methods to characterize and model scatter in SPECT and correct for its image degrading effects, both for clinical and small animal SPECT. Traditionally scatter correction methods were limited in accuracy, noise properties and/or generality and were not very widely applied. For small animal SPECT, these approximate methods of correction are often sufficient since the fraction of detected scattered photons is small. This contrasts with patient imaging where better accuracy can lead to significant improvement of image quality. As a result, over the last two decades, several new and improved scatter correction methods have been developed, although often at the cost of increased complexity and computation time. In concert with (i) the increasing number of energy windows on modern SPECT systems and (ii) excellent attenuation maps provided in SPECT/CT, some of these methods give new opportunities to remove degrading effects of scatter in both standard and complex situations and therefore are a gateway to highly quantitative single- and multi-tracer molecular imaging with improved noise properties. Widespread implementation of such scatter correction methods, however, still requires significant effort.     

Cell-Mediated Resistance Induced with Immunogenic Preparations of Salmonella typhimurium

Peritoneal cells obtained from mice immunized 15 or 30 days previously with (i) 0.1 LD(50) of attenuated Salmonella typhimurium (RIA), (ii) 20 mug (dry weight) of heat-killed Salmonella (SR-11), (iii) 20 mug (dry weight) of immunogenic ribosomal subfractions, or (iv) 20 mug of ribonucleic acid (RNA) subfractions were passively transferred to normal unimmunized mice. The ability of the recipient animals to inhibit or retard the multiplication of virulent challenge S. typhimurium 5 days post-infection was determined by pathogen counts on the carcasses. Peritoneal cells from donors immunized with the RIA, ribosomal, or RNA preparations (i) conferred maximal resistance to challenge 10 to 15 days after cell transfer and demonstrable resistance throughout the 45-day assay period, (ii) conferred resistance to infection when 10(5), 10(3), or 10(2) peritoneal cells were injected subcutaneously but not with fewer than 10(5) cells administered intraperitoneally, and (iii) rendered recipients capable of acting as donors of peritoneal cells that conferred demonstrable resistance on normal recipients. Recipients of peritoneal cells obtained from donors immunized with heat-killed bacteria were unable to (i) significantly inhibit bacterial proliferation at 10 days post-transfer, (ii) resist infection by a challenge inoculum greater than 50 LD(50), and (iii) secondarily confer resistance on normal animals through the passive transfer of cells or serum.     

Apoptosis of ejaculated human sperm is induced by co-incubation with Chlamydia trachomatis lipopolysaccharide

BACKGROUND: Previous work has shown that co-incubation of human sperm with Chlamydia trachomatis serovars E and LGV leads to premature sperm death and that this is due primarily to chlamydial lipopolysaccharide (LPS). Here, we investigated the possible involvement of apoptosis in this premature sperm death. METHODS: Highly motile preparations of sperm from normozoospermic patients were co-incubated for 6 h with extracted LPS from C. trachomatis serovars E and LGV. Three different methods were used to determine if LPS-treated sperm underwent apoptosis, including: (i) flow cytometry; (ii) measurement of ADP:ATP ratios; and (iii) measurement of mono- and oligonucleosomal DNA fragments. Caspase activity was also investigated by fluorimetry and by use of a pan-caspase inhibitor and caspase-3 inhibitor. RESULTS: All three methods used for detection indicated that C. trachomatis LPS induced some apoptosis in sperm after 6 h when compared with a staurosporine (apoptosis-positive) control. Moreover, a greater degree of apoptosis was seen with C. trachomatis serovar E than with serovar LGV. It was also shown that C. trachomatis LPS-induced apoptosis of sperm could be blocked with a pan-caspase inhibitor and a caspase-3 inhibitor. Moreover, by using a fluorogenic substrate, apoptosis was shown to be caspase-mediated. CONCLUSIONS: In general it is believed that apoptosis does not occur in C. trachomatis-infected host cells. However, using three different methods, our findings clearly indicate that co-incubation of sperm with C. trachomatis LPS results in cellular death which is in part due to apoptosis and is caspase-mediated. These findings provide an explanation as to how C. trachomatis can mediate premature death in human sperm.     

The effect of cooking methods on the allergenicity of peanut

It has been rarely studied why the prevalence of peanut allergy is relative low in China. The aim of this study was to investigate: (i) the major peanut allergens in China, (ii) the effect of buffer composition on peanut allergens extraction efficiency, and (iii) the effect of cooking methods on peanut allergenicity. The allergenic property of peanut protein extracts was assessed by immunoblotting. The relative contents of the major peanut allergens were quantified by densitometry. The ability of IgE to bind to the different peanut preparations (fried, boiled and roasted) was detected with ELISA. In conclusion, Ara h 1 and several proteins belonging to the Ara h 3 were major peanut allergens. Different buffers lead to different extraction efficiency for allergen fragments. The IgE binding property of peanuts processed by cooking, frying and roasting did not differ significantly. Therefore, cooking methods may not explain the reason for the lower prevalence of peanut allergy in China.     

Non-genomic effects of glucocorticoids in the neural system. Evidence, mechanisms and implications

Complementing the classical concept of genomic steroid actions, here we (i) review evidence showing that important neural effects of glucocorticoids are exerted by non-genomic mechanisms; (ii) describe known mechanisms that may underlie such effects; (iii) summarize the functions and implications of non-genomic mechanisms and (iv) outline future directions of research. The role of non-genomic mechanisms is to shape the response of the organism to challenges that require a substantial reorganization of neural and somatic functions and involve massive behavioral shifts. Non-genomic effects may (i) prepare the cell for subsequent glucocorticoid-induced genomic changes, (ii) bridge the gap between the early need of change and the delay in the expression of genomic effects and (iii) may induce specific changes that in some instances are opposite to those induced by genomic mechanisms. The latter can be explained by the fact that challenging situations require different responses in early (acute) and later (chronic) phases. Data show that non-genomic mechanisms of glucocorticoid action play a role in both pathological phenomena and the expression of ameliorative pharmacological effects. Non-genomic mechanisms that underlie many glucocorticoid-induced neural changes constitute a for long overlooked controlling factor. Despite the multitude and the variety of accumulated data, important questions remain to be answered.     

Chemosensitivity of medullary neurons in explant tissue cultures.

To determine whether cultured medulla contains chemosensitive neurons which are excited by CO2 and fixed acid and whether this function is specific to the ventral medulla, tissue explants of ventral and dorsal medulla were prepared from neonatal rats and incubated for two to three weeks. Cultures were superfused with artificial cerebrospinal fluid, maintained at 37 degrees C, and pH of the superfusate was varied either with PCO2 (14-71 Torr) at constant HCO3- (22 mM) or HCO3- (10-30 mM) at constant PCO2 (35 Torr). Spontaneous action potentials were recorded extracellularly in 51 ventral and 23 dorsal medullary neurons. Ventral medullary neurons exhibited a steady baseline firing frequency of 4 +/- 0.8 Hz. In contrast, dorsal medullary neurons exhibited two different patterns of spontaneous activity: 11 fired continuously (7.2 +/- 1.4 Hz) while 12 fired with a bursting pattern. Burst duration was 0.80 +/- 0.14 min and cycle time was 1.74 +/- 0.43 min. Decreasing pH with CO2 caused an increase in the activity of 10 of 27 ventral medullary neurons and two of six dorsal medullary neurons with a mean response of 7.5 Hz/pH unit. Varying pH by changing HCO3- had no effect on firing frequency. These results demonstrate that: (i) chemosensitive neurons are present in both ventral and dorsal medullary explant cultures; (ii) these cells only respond to changes in pH induced with CO2; and (iii) about half of the dorsal medullary neurons fire spontaneously with a regular bursting pattern of activity.     

Comparison of methods for determination of the number of active centers in Ziegler-Natta polymerizations

Four intrinsically different methods for the determination of the number of active centers were employed in propylene polymerization catalyzed by the typical heterogeneous system TiCl₃/AlEt₂Cl: (i) determination of M?_n (GPC method) and calculation of number of macromolecules therefrom; (ii) quenching with BuOT and determination of tritium in the polymer; (iii) quenching with CO or CO₂ and determination of carbonyl groups in the polymer (IR spectroscopy); (iv) quenching with SO₂ and determination of sulfur in the polymer. The first two methods exhibited a reasonable agreement for the number of active centers. Method (iii) gave much lower values, whereas method (iv) was found to be virtually unsuitable due to a high extent of side reactions of SO₂ with the polymer. A similar comparison of the methods was made in ethylene polymerization catalyzed by the typical homogeneous systems Cp₂TiEtCl/AlEtCl₂ and Cp₂TiEtCl/alumoxane. Method (ii) gave a much lower number of active centers than method (i), due perhaps to a low extent of alcoholysis of the Ti? C bond or to a high kinetic isotope effect. CO₂ appeared to be entirely inert as a quenching agent and CO had been reported to show side reactions. It follows that none of the above quenching methods have universal application and thus literature data should be considered with caution.     

Synaptic actions of peripheral nerve impulses upon Deiters neurones via the mossy fibre afferents

1. The cerebellar integration of sensory inputs to Deiters neurones was investigated in decerebrate cats. In some preparations decerebration was combined with transection of the olivocerebellar fibres.2. In the latter preparations peripheral nerve impulses generally produced a response consisting of a sequence of the following post-synaptic potentials: (i) an initial e.p.s.p. (d(1)), (ii) early i.p.s.p. (h(1)), (iii) later i.p.s.p. (h(2)).3. The mean latencies of d(1), h(1) and h(2) were 5.7, 7.3 and 9.8 msec from the forelimb nerves, and 7.5, 9.0 and 13.4 msec from the hind limb nerves, respectively.4. The stimulus intensity-response relation indicates that the Group I muscle afferents as well as the low threshold cutaneous afferents contribute to the response.5. In the preparations with the intact inferior olive there were additional components of the post-synaptic potentials: a later e.p.s.p. (d(2)) and another later i.p.s.p. (h(3)), their mean latencies being 15.3 and 19.7 msec from the forelimb nerves, and 18.0 and 21.3 msec from the hind limb nerves, respectively.6. The d(1) and h(2) components were attributed to the mossy fibre afferents and d(2) and h(3) to the climbing fibres; d(1) and d(2) were due to excitation through the collaterals of the mossy and climbing fibres, and h(2) and h(3) to inhibition from Purkyne cells activated by the mossy and climbing fibres, respectively. h(1) was too early to be produced through the cerebellum, and was probably mediated by inhibitory neurones in the reticular formation.