Reading between the spikes of the hypothalamic neural code.

Reading the spike coding of hypothalamic neurones presents a considerable challenge because they exhibit highly irregular firing patterns. Electrophysiologists working in the motor and sensory systems, in which neurones fire more regularly, have devised satisfactory methods to describe the firing of cells, although the statistical assumptions that underlie the methods do not apply to hypothalamic neurones. Measurement of neural activity is nevertheless vital to characterise the activity of neuroendocrine cells. It has thus become necessary to develop methods suitable for the analysis of the highly irregular spike discharge patterns of both spontaneous and stimulus-evoked firing of hypothalamic neurones. We review techniques used to meet this challenge and demonstrate their considerable capacity to address important physiological questions. We also introduce a novel approach for valid statistical estimation of the information conveyed by the response of a single neurone to a periodic stimulus. The approach demonstrated significant diurnal rhythms of synaptic connectivity between hypothalamic nuclei.

Daily rhythms of spike coding in the rat supraoptic nucleus.

Novel measures of coding based on interspike intervals were used to characterise the rhythms of single unit activity in the supraoptic nucleus during the day/night cycle in urethane-anaesthetised rats in vivo. Both continuously firing and phasic cells showed significant (P < 0.001) diurnal rhythms of spike frequency and in the irregularity of firing, as quantified by the log interval entropy (ENT). Comparison of rhythms in log interval ENT showed that the amplitude of the rhythms was greater for the continuously firing cells than for the phasic cells (P = 0.002). Rhythms persisted after hypertonic stimulation or pinealectomy and both treatments reduced the amplitude significantly only for the continuously firing cell group. By contrast, the mesor (i.e. mid-point of the rhythm) was reduced only for the phasic cell group. A similar analysis applied to the activity of cells of the suprachiasmatic nucleus showed that, after pinealectomy, there was a significant rhythm in ENT (P < 0.001) but not firing rate; however, the amplitude of the rhythm in ENT was attenuated (P = 0.047). Diurnal changes in the electrical activity of supraoptic cells are consistent with previously reported circadian changes in magnocellular neuropeptide release. Differences between continuous and phasic cell groups in the effects of osmotic stimulation on rhythmic activity indicate that the two cell types differ in their coding of osmolality and zeitgeber time information. The different effects of pinealectomy on the supraoptic and suprachiasmatic nuclei suggest that removal of endogenous melatonin unmasks a difference in circadian coding between the two nuclei.

Secretory cells of the supraoptic nucleus have central as well as neurohypophysial projections.

Conventional neuroanatomical methods may fail to demonstrate the presence of axons that are finer than 1 microm in diameter because such processes are near or below the limit of resolution of the light microscope. The presence of such axons can, however, be readily demonstrated by recording. The most easily interpreted type of recording for this purpose is the demonstration of antidromic activation of the cell body following stimulation of the region through which the axon passes. We have exploited this technique in the hypothalamus and have demonstrated the presence of double axonal projections or axons branching very near the cell bodies of the secretory cells of the neurohypophysial system in the rat supraoptic nucleus. We found that a small proportion of supraoptic magnocellular cells could be antidromically activated both from the neural stalk and from elsewhere in the hypothalamus, including the suprachiasmatic nucleus (8 cells of a total of 182) and the antero-ventral third ventricular region (AV3V; 4 of 182 cells) near the organum vasculosum of the lamina terminalis (OVLT). Collision of antidromic and orthodromic spikes showed that the cells were clearly antidromically (rather than synaptically, or orthodromically) activated from both sites. A stimulus applied to one of the axons prevented propagation of a spike evoked by a pulse delivered to the other axon until sufficient time had elapsed after the first stimulus for the resultant spike to have propagated from the first stimulus site along one cell process (towards the cell body or branch point), and from this point along the other axonal branch to the second stimulus site (there was also a short additional delay period during which the axon at the site of the second stimulus recovered from its absolute refractory period). If the interval between the stimuli was progressively reduced, there came a point where the second spike failed. Such a clear demonstration of dual projections in a system where the cells were previously thought to have only a single axon raises the possibility that many nerve cells in the CNS have previously unsuspected projections.

Leptin modulates spike coding in the rat suprachiasmatic nucleus.

The mammalian circadian pacemaker, the suprachiasmatic nucleus (SCN), contains receptors to the adipose tissue hormone leptin. In the present study, the effects of leptin on the electrophysiological activity of the SCN cells were characterised in vitro in rat brain slices. During extracellular recording, application of 20 nm leptin (n = 36) decreased mean spike frequency (Wilcoxon signed rank test, z = -3.390, P < 0.001) and increased the irregularity of firing measured by the entropy of the log interspike interval distribution (Student's paired t-test, t = 2.377, P = 0.023), but had no consistent effect on spike patterning as measured by the mutual information between adjacent log interspike intervals (z = 0.745, P = 0.456). Intracellular current-clamp recordings (n = 25) revealed a hyperpolarising effect of 20 nm leptin on SCN neurones (z = -2.290, P = 0.022). The hyperpolarisation largely resulted from the effect of leptin on the subgroup of cells (n = 13) that generated 'rebound' spikes upon termination of a hyperpolarising current pulse (z = -2.697, P = 0.007). Leptin application also increased the group mean duration of the afterhyperpolarisation (n = 25, t = 2.512, P = 0.023). The effects of leptin on extracellularly recorded spike activity were consistent with the changes in membrane potential and spike shape. They suggest that leptin can directly modulate the electrical properties of SCN neurones and, in this way, contribute to the mechanism by which metabolic processes influence the circadian clock.

Osmotic modulation of stimulus-evoked responses in the rat supraoptic nucleus.

Neural information is conveyed by action potentials along axons to downstream synaptic targets. Synapses permit functionally relevant modulation of the information transmitted by converging inputs. Previous studies have measured the amount of information associated with a given stimulus based either on spike counts or on the relative frequencies of spike sequences represented as binary strings. Here we apply information theory to the phase-interval stimulus histogram (PhISH) to measure the extent of the stimulus-evoked response using the statistical relationship between each interspike interval and its phase within the stimulus cycle. We used the PhISH as a novel approach to investigate how different osmotic states affect the flow of information through the osmoreceptor complex of the hypothalamus. The amount of information conveyed from one (afferent) element of the complex, the anteroventral region of the third ventricle (AV3V), to another (an efferent element), the supraoptic nucleus, was increased by hypertonic stimulation (intravenous mannitol, z = 4.39, P < 0.001) and decreased by hypotonic stimulation (intragastric water, z = -3.37, P < 0.001). Supraoptic responses to AV3V stimulation differed from those that follow stimulation of a hypothalamic element outside the osmoreceptor complex, the suprachiasmatic nucleus (SCN), which also projects to the supraoptic nucleus. Thus osmosensitive gain control mechanisms differentially modulate osmotically dependent and osmotically independent inputs, and enhance the osmoresponsiveness of supraoptic cells within a physiological range. The value of the novel approach is that its use is not limited to the osmoreceptor ensemble but it can be used to investigate the flow of information throughout the central nervous system.

Melatonin modulates spike coding in the rat suprachiasmatic nucleus.

The effects of the application of melatonin in vitro on the electrophysiological activity of suprachiasmatic neurones were characterised using novel measures of coding based on the analysis of interspike intervals. Perfusion of 1 nM melatonin in vitro (n = 53) had no consistent effect on mean spike frequency (Wilcoxon's sign rank, z = -0.01, P = 0.989), but increased the irregularity of firing (Student's paired t-test, t = -3.02, P = 0.004), as measured by the log interval entropy, and spike patterning (z = -3.43, P < 0.001), as measured by the mutual information between adjacent log intervals. Intracellular recordings in vitro in current clamp mode showed that 1 nM melatonin significantly hyperpolarised (n = 11, z = -2.35, P = 0.019) those cells that showed 'rebound' spikes upon termination of a hyperpolarising current pulse. Grouping all cells together (n = 27), melatonin application decreased the duration of the afterhyperpolarisation (z = -2.49, P = 0.013) and increased the amplitude of the depolarising afterpotential (z = -2.71, P = 0.007). The effects of melatonin seen in vitro from extracellular recordings on interspike interval coding were consistent with the changes in spike shape seen from intracellular recordings. A melatonin-induced increase in the size of the depolarising afterpotential of suprachiasmatic cells might underlie the increased irregularity of spike firing seen during the subjective night time. The method of analysis demonstrated a difference in spike firing that is not revealed by frequency alone and is consistent with the presence of a melatonin-induced depolarising current.

Pinealectomy reduces optic nerve but not intergeniculate leaflet input to the suprachiasmatic nucleus at night.

The suprachiasmatic nucleus (SCN) of the hypothalamus regulates circadian rhythms in mammals. It receives, among others, direct inputs from the retina and from the thalamic intergeniculate leaflet (IGL). The former sends photic signals to the SCN, whereas the latter probably integrates photic and nonphotic information. To characterise these inputs in vivo, extracellular single-unit recordings were made from the SCN of rats under urethane anaesthesia during electrical stimulation of the optic nerve (OptN) or the IGL region. Cell responses were evaluated by creating peri-stimulus time histograms. Because humoral signals such as melatonin might modulate the activity of the SCN in addition to neural inputs, recordings were also made using pinealectomised (Px) rats to test for a possible role of this hormone in regulating inputs to the SCN. A significantly greater number of cells responded to IGL (60 of 90, 67%) than to OptN (35 of 75, 47%) stimulation in intact animals (chi(2) = 5.905, P = 0.015). The same was true when Px animals were tested (IGL, 82 of 131, 63%; OptN, 31 of 111, 28%; chi(2) = 27.637, P < 0.001). In intact animals, the proportion of cells responsive to IGL stimulation during the day and during the night was not significantly different from the proportion responsive in Px animals. The same was true for OptN stimulation during the day. However, during the night, the proportion of cells responsive to OptN stimulation in intact animals was significantly greater than the proportion responsive in Px animals (chi(2) = 7.127, P = 0.008). Our findings suggest that a lack of melatonin modulates OptN but not IGL inputs to the SCN.

Spike coding during osmotic stimulation of the rat supraoptic nucleus.

Novel measures of coding based on interspike intervals were used to characterize the responses of supraoptic cells to osmotic stimulation. Infusion of hypertonic NaCl in vivo increased the firing rate of continuous (putative oxytocin) cells (Wilcoxon z= 3.84, P= 0.001) and phasic (putative vasopressin) cells (z= 2.14, P= 0.032). The irregularity of activity, quantified by the log interval entropy, was decreased for continuous (Student's t= 3.06, P= 0.003) but not phasic cells (t= 1.34, P= 0.181). For continuous cells, the increase in frequency and decrease in entropy was significantly greater (t= 2.61, P= 0.036 and t= 3.06, P= 0.007, respectively) than for phasic cells. Spike patterning, quantified using the mutual information between intervals, was decreased for phasic (z=-2.64, P= 0.008) but not continuous cells (z=-1.14, P= 0.256). Although continuous cells showed similar osmotic responses to mannitol infusion, phasic cells showed differences: spike frequency decreased (z=-3.70, P < 0.001) and entropy increased (t=-3.41, P < 0.001). Considering both cell types together, osmotic stimulation in vitro using 40 mm NaCl had little effect on firing rate (z=-0.319, P= 0.750), but increased both entropy (t= 2.75, P= 0.010) and mutual information (z=-2.73, P= 0.006) in contrast to the decreases (t= 2.92, P= 0.004 and z=-2.40, P= 0.017) seen in vivo. Responses to less severe osmotic stimulation with NaCl or mannitol were not significant. Potassium-induced depolarization in vitro increased firing rate (r= 0.195, P= 0.034), but the correlation with decreased entropy was not significant (r=-0.097, P= 0.412). Intracellular recordings showed a small depolarization and decrease in input resistance during osmotic stimulation with NaCl or mannitol, and membrane depolarization following addition of potassium. Differences in responses of oxytocin and vasopressin cells in vivo, suggest differences in the balance between the synaptic and membrane properties involved in coding their osmotic responses. The osmotic responses in vivo constrasted with those seen in vitro, which suggests that, in vivo, they depend on extrinsic circuitry. Differences in responses to osmolality and direct depolarization in vitro indicate that the mechanism of osmoresponsiveness within a physiological range is unlikely to be fully explained by depolarization.

Spike coding from the perspective of a neurone.

In this paper, we compare existing methods for quantifying the coding capacity of a spike train, and review recent developments in the application of information theory to neural coding. We present novel methods for characterising single-unit activity based on the perspective of a downstream neurone and propose a simple yet universally applicable framework to characterise the order of complexity of neural coding by single units. We establish four orders of complexity in the capacity for neural coding. First-order coding, quantified by firing rates, is conveyed by frequencies and is thus entirely described by first moment processes. Second-order coding, represented by the variability of interspike intervals, is quantified by the log interval entropy. Third-order coding is the result of spike motifs that associate adjacent inter-spike intervals beyond chance levels; it is described by the joint interval histogram, and is measured by the mutual information between adjacent log intervals. Finally, nonstationarities in activity represent coding of the fourth-order that arise from the effects of a known or unknown stimulus.

Rhythmic changes in spike coding in the rat suprachiasmatic nucleus.

The suprachiasmatic nucleus is regarded as the main mammalian circadian pacemaker but evidence for rhythmic firing of single units in vivo has been obtained only recently. The present study was undertaken to determine if rhythms could be seen using measures of activity in addition to the mean spike frequency. We investigated whether there were changes in the irregularity of cell activity measured by the disorder of the interspike interval distribution for neurones recorded in vivo and in vitro. By plotting the entropy of the log interval histogram that quantifies the coding capacity for each action potential against the respective zeitgeber time, we describe oscillations of spike activity in vivo. Entropy measures have the advantage over variances in that they quantify aspects of the shape of the distribution and not just the dispersion. One hundred and sixty-six cell recordings from the suprachiasmatic nucleus showed a significant rhythm in entropy with an oscillatory trend in the data (P < 0.001) showing a trough towards the end of the light period and a peak in the mid-dark period. There was a similar rhythm for the cells recorded from the peripheral zone (n = 209, P = 0.037). In separate experiments in vitro, to investigate the relationship between mean spike frequency and entropy, potassium-induced depolarization of cells recorded during the subjective night was correlated with a significant increase in mean spike frequency (r = 0.259, P = 0.011) and a decrease in entropy (r = -0.296, P = 0.004). The negative correlation between the entropy and mean spike frequency of cells recorded in vitro was significantly different from that seen in vivo (F = 15.5, P < 0.001), which may reflect differences in the balance between deterministic and stochastic influences on spike occurrence. The study shows that while there is a rhythm of mean spike frequency, parameters based on the variability of interspike interval distributions also display rhythmic changes over the day-night cycle.

Assessment of spike activity in the supraoptic nucleus.

Novel approaches to the characterization of coding carried by spike trains are discussed. Measuring firing frequency alone may only partially reflect spike patterning, and can only quantify changes of the most obvious kind. We have devised a method that combines probabilistic and information approaches to quantify the variability of the interspike intervals in a way that is independent of spike frequency. To illustrate the technique, the firing of an oxytocin cell and a vasopressin cell were compared before and after osmotic stimulation. A bimodal lognormal function was fitted to the interspike interval histograms. The entropy of the log interval histogram was used to measure the variability of intervals and to reflect the coding capacity of the cell per spike. A perfect metronome shows no variability in interval and thus has no greater coding capacity than is conveyed by its frequency, whereas the variability of intervals of magnocellular neurones means that their irregular activity has greater potential for coding. While the mean spike frequency increased in both the oxytocin and vasopressin cells in response to osmotic stimulation, the changes in their irregularity showed differences. Osmotic stimulation reduced the entropy of the oxytocin cell, reflecting an increase in the regularity of its spike activity. Conversely, osmotic stimulation had little effect on the entropy of the vasopressin cell. Such differences are not evident from a simple inspection of ratemeter activity. The comparison highlights the limitations of mean spike frequency as a measure of spike coding. Parameters based on the interspike intervals constitute informative measures of spike activity that allow objective comparisons to be made between the activity under different physiological conditions.

Responses of cells in the rat supraoptic nucleus in vivo to stimulation of afferent pathways are different at different times of the light/dark cycle.

To determine whether the daily rhythms of spike activity in the supraoptic nucleus (SON) were accompanied by changes in the behaviour of its inputs, we used conventional extracellular single cell recordings from cells in the SON of anaesthetized rats while stimulating the contralateral optic nerve and the ipsilateral suprachiasmatic nucleus (SCN). Neurones in the SON region were identified by antidromic activation and classified as oxytocin or vasopressin cells, on the basis of their spontaneous firing patterns. Approximately 27% of both oxytocin (29/108) and vasopressin (39/147) neurones were excited by stimulation of the optic nerve, and the majority of responses had a long latency (>20 ms). Very few oxytocin (3/108) and vasopressin cells (2/147) were inhibited by stimulation of the optic nerve. The pattern of the responses (excitatory, inhibitory or nonresponsive) of oxytocin and vasopressin cells to stimulation of the optic nerve was significantly related to the time of day (chi-square test; P = 0.012, oxytocin cells; P = 0.006, vasopressin cells). The proportion of oxytocin cells excited by stimulation of the optic nerve was highest at ZT 4-8 and lowest at ZT 20-24. For vasopressin cells, it was highest at ZT 12-16 and lowest at ZT 20-24. The proportion of excitatory, inhibitory and complex responses seen in oxytocin and vasopressin cells following stimulation of the SCN also changed and was significantly different at different times of day (oxytocin cells: highest proportion of excitatory responses at ZT 12-16, P = 0.029; chi-square test; vasopressin cells: highest proportion of excitatory responses at ZT 0-4, P = 0.005; chi-square test). Thus, inputs to oxytocin and vasopressin neurones from the optic nerve and some outputs from the SCN changed during the light/dark cycle. Such changes may contribute to the generation of 24-h rhythms in activity of oxytocin and vasopressin neurones and release of the peptides.

Measuring spike coding in the rat supraoptic nucleus.

Measuring spike coding objectively is essential to establish whether activity recorded under one set of conditions is truly different from that recorded under another set of conditions. However, there is no generally accepted method for making such comparisons. Measuring firing frequency alone only partially reflects spike patterning. In this paper, novel quantities based on the logarithmic interspike intervals are proposed as useful measures of spontaneous activity. We illustrate the methods by comparing extracellular recordings from magnocellular cells of the rat supraoptic nucleus in vivo and in vitro and between oxytocin and vasopressin cells in vivo. A bimodal Gaussian function fitted to the log interspike interval histogram accurately described the distribution profile for very different types of activity. We introduce the entropy of the log interval distribution as a novel quantity that measures the capacity of a cell to encode information other than a constant instantaneous frequency. Unlike existing entropy measures that are based on spike counts, it quantifies the variability in the interval distribution. In addition, the mutual information between adjacent log intervals is proposed as an objective measure of patterned activity. For cells recorded in vivo and in vitro, there was no significant difference in mean spike frequencies but there were differences in the log interval entropy (t = -4.97, P < 0.001) and the mutual information (z = -2.64, P < 0.01). The differences may result from the disruption of connections in the slice preparation. When a comparison was made between the spike activity of oxytocin and vasopressin cells recorded in vivo, there was a difference in mutual information (z = 5.15, P < 0.001) but not in mean spike frequency. Both comparisons highlight the potential limitations of using mean spike frequency alone as a measure of spike coding. We propose that our novel parameters based on interval analysis constitute informative measures of spontaneous activity under different physiological conditions.

Responses of cells in the rat suprachiasmatic nucleus in vivo to stimulation of afferent pathways are different at different times of the light/dark cycle.

Conventional extracellular recordings were made from single cells in the suprachiasmatic nucleus (SCN) region of the anaesthetized rat. Each cell was tested for its response to stimulation at three sites; the contralateral optic nerve, the ipsilateral supraoptic nucleus (SON) or the ipsilateral arcuate nucleus (ARC) to determine whether the behaviour of the synapses in the SCN was different at different times. Responses to stimulation were tested once each hour and assessed by creating peristimulus time histograms. Excitatory, inhibitory or complex (consisting of more than one component) responses were seen. The responses of some cells that were recorded for several hours changed with time. Changes were seen in the responses of SCN cells to stimulation of the ARC (31/91 cells) and the SON (26/90 cells) regions, but only rarely to stimulation of the optic nerve (2/72 cells). Such differences in proportion are unlikely to have occurred by chance (P < 0.001; chi-square test). Changes seen included the appearance of both excitatory and inhibitory responses in cells that were initially unresponsive. In some cells, one component of a complex response remained constant while another component changed with time. When the cells in the SCN were treated as a group, the proportion of excitatory, inhibitory or complex responses to ARC stimulation did not remain constant throughout the light/dark cycle (P = 0.014; chi-square test). The proportion of excitatory, inhibitory or complex responses to SON and optic nerve stimulation showed no significant variation with the light/dark cycle. If a change in response can be interpreted as a change in the behaviour of a neural connection, the results imply that some of the projections to the SCN from within the hypothalamus change at different times of the light/dark cycle, whereas no change could be seen in the input from the optic nerve. Thus, some of the connections of the SCN appear not to be hard wired, but change rapidly with time.

Defined cell groups in the rat suprachiasmatic nucleus have different day/night rhythms of single-unit activity in vivo.

The electrical activity of the rat suprachiasmatic nucleus (SCN) was examined in anesthetized rats in vivo using single-unit electrophysiological techniques. The present data confirm the daily variation in the electrical activity of the SCN previously reported in vitro and in vivo using multiple-unit recording techniques. They further suggest that subpopulations of suprachiasmatic neurons with different neural connections have a different daily rhythm of activity. Neurons in the SCN region showed a significant rhythm of activity (p = 0.034; Kruskall-Wallis analysis of variance [KW-ANOVA]). The greatest activity occurred during the second part of the light period (ZT 10-12), and the lowest activity occurred in the early part of the light period (ZT 0-2). The subgroup of cells in the suprachiasmatic region with output projections to the arcuate nucleus (ARC) and/or supraoptic nucleus (SON) regions also showed a significant rhythm (p = 0.001; K-W ANOVA). Their activity appeared to show two peaks near the light-dark (ZT 10-12) and dark-light (ZT 22-24) transition periods with the lowest activity at ZT 16-18. This rhythm was significantly different (p = 0.016) from that of neurons without an output projection to the ARC and/or SON. Retinorecipient suprachiasmatic neurons appeared to have a less robust daily rhythm in their activity. The change in the firing behavior of the cells was not reflected simply by changes in mean firing rate. Examination of the coefficient of variation of the interspike interval distribution of cells at different times of day revealed changes in the firing pattern of cells in the SCN region that did not have output projections (p = 0.032; K-W ANOVA). The present results thus suggest that the SCN is composed of a heterogeneous population of neurons and that different rhythms of activity are expressed by neurons with different neural connections. There were changes in both firing pattern and firing rate.