Extracting Information from the Power Spectrum of Synaptic Noise

In cortical neurons, synaptic "noise" is caused by the nearly random release of thousands of synapses. Few methods are presently available to analyze synaptic noise and deduce properties of the underlying synaptic inputs. We focus here on the power spectral density (PSD) of several models of synaptic noise. We examine different classes of analytically solvable kinetic models for synaptic currents, such as the "delta kinetic models," which use Dirac delta functions to represent the activation of the ion channel. We first show that, for this class of kinetic models, one can obtain an analytic expression for the PSD of the total synaptic conductance and derive equivalent stochastic models with only a few variables. This yields a method for constraining models of synaptic currents by analyzing voltage-clamp recordings of synaptic noise. Second, we show that a similar approach can be followed for the PSD of the the membrane potential (Vm) through an effective-leak approximation. Third, we show that this approach is also valid for inputs distributed in dendrites. In this case, the frequency scaling of the Vm PSD is preserved, suggesting that this approach may be applied to intracellular recordings of real neurons. In conclusion, using simple mathematical tools, we show that Vm recordings can be used to constrain kinetic models of synaptic currents, as well as to estimate equivalent stochastic models. This approach, therefore, provides a direct link between intracellular recordings in vivo and the design of models consistent with the dynamics and spectral structure of synaptic noise.     

Dipole characterization of single neurons from their extracellular action potentials

The spatial variation of the extracellular action potentials (EAP) of a single neuron contains information about the size and location of the dominant current source of its action potential generator, which is typically in the vicinity of the soma. Using this dependence in reverse in a three-component realistic probe + brain + source model, we solved the inverse problem of characterizing the equivalent current source of an isolated neuron from the EAP data sampled by an extracellular probe at multiple independent recording locations. We used a dipole for the model source because there is extensive evidence it accurately captures the spatial roll-off of the EAP amplitude, and because, as we show, dipole localization, beyond a minimum cell-probe distance, is a more accurate alternative to approaches based on monopole source models. Dipole characterization is separable into a linear dipole moment optimization where the dipole location is fixed, and a second, nonlinear, global optimization of the source location. We solved the linear optimization on a discrete grid via the lead fields of the probe, which can be calculated for any realistic probe + brain model by the finite element method. The global source location was optimized by means of Tikhonov regularization that jointly minimizes model error and dipole size. The particular strategy chosen reflects the fact that the dipole model is used in the near field, in contrast to the typical prior applications of dipole models to EKG and EEG source analysis. We applied dipole localization to data collected with stepped tetrodes whose detailed geometry was measured via scanning electron microscopy. The optimal dipole could account for 96% of the power in the spatial variation of the EAP amplitude. Among various model error contributions to the residual, we address especially the error in probe geometry, and the extent to which it biases estimates of dipole parameters. This dipole characterization method can be applied to any recording technique that has the capabilities of taking multiple independent measurements of the same single units.     

Spatial coherence and stationarity of local field potentials in an isolated whole hippocampal preparation <Emphasis Type="Italic">in vitro</Emphasis>

Local field potential (LFP) multielectrode recordings of spontaneous rhythms in an isolated whole hippocampal preparation are characterized with respect to their spatial variability within the hippocampus, and their frequency properties. Using simulated data, we categorize potential relationships between frequency variation over time in LFP recordings and spatial variability between electrodes. We then use data recorded from the intact preparation to distinguish between our theoretical categories. We find that the LFP recordings have a close to spatially invariant frequency distribution (not phase) across the hippocampus, and differ in frequency only in a component that may be seen as physiological noise. From these facts, we conclude that the isolated hippocampal LFP recordings represent a single signal and may be regarded as a unitary circuitry. We additionally examine phase differences across our recording sites. We use our characterization of the hippocampal isolate’s properties to predict its spatial coherence in response to high frequency stimulation. We find that there is a finely tuned inverse relationship between temporal variability in the hippocampal isolate’s LFP recordings and their spatial coherence.     

Interpretation of EMG changes with fatigue: facts, pitfalls, and fallacies.

Failure to maintain the required or expected force, defined as muscle fatigue, is accompanied by changes in muscle electrical activity. Although studied for a long time, reasons for EMG changes in time and frequency domain have not been clear until now. Many authors considered that theory predicted linear relation between the characteristic frequencies and muscle fibre propagation velocity (MFPV), irrespective of the fact that spectral characteristics can drop even without any changes in MFPV, or in proportion exceeding the MFPV changes. The amplitude changes seem to be more complicated and contradictory since data on increased, almost unchanged, and decreased amplitude characteristics of the EMG, M-wave or motor unit potential (MUP) during fatigue can be found in literature. Moreover, simultaneous decrease and increase in amplitude of MUP and M-wave, detected with indwelling and surface electrodes, were referred to as paradoxical. In spite of this, EMG amplitude characteristics are predominantly used when causes for fatigue are analysed. We aimed to demonstrate theoretical grounds for pitfalls and fallacies in analysis of experimental results if changes in intracellular action potential (IAP), i.e. in peripheral factors of muscle fatigue, were not taken into consideration. We based on convolution model of potentials produced by a motor unit and detected by a point or rectangular plate electrode in a homogeneous anisotropic infinite volume conductor. Presentation of MUP in the convolution form gave us a chance to consider power spectrum (PS) of MUP as a product of two terms. The first one, PS of the input signal, represented PS of the first temporal derivative of intracellular action potential (IAP). The second term, PS of the impulse response, took into account MFPV, differences in instants of activation of each fibre, MU anatomy, and MU position in the volume conductor in respect to the detecting electrode. PS presentation through product means that not only changes in MFPV could be responsible for PS shift as is usually assumed. Changes in IAP duration and IAP after-potential magnitude, affecting the first term of the product, influence the product and thus MUP PS. Moreover, the interrelations between the two spectra and thus sensitivity of spectrum to different parameters change with MU-electrode distance because the second term depends on it. Thus, we have demonstrated that theory does not predict a linear relation between the characteristic frequencies (maximum, mean and median) and MFPV. IAP duration and after-potential magnitude are among parameters affecting MUP or M-wave PS and thus, EMG PS detected by monopolar and bipolar electrodes. Usage of single fibre action potential models instead of MUP ones can result in false dependencies of frequency characteristics. The MUP amplitude characteristics are determined not only by amplitude of IAP, but also by the length of the IAP profile and source-electrode distance. Due to the IAP profile lengthening and an increase in the negative after-potential, surface detected EMG amplitude characteristics can increase even when IAP amplitude decreases considerably during fatigue. Increase in surface detected MUP or M-wave amplitude should not be attributed to a weaker attenuation of the low-frequency components with distance. Simultaneous decrease and increase in amplitude of MUP and M-wave detected with indwelling and surface electrodes are regular, not paradoxical. Corner frequency of the high pass filter should be 0.5 or 1 Hz when muscle fatigue is analyzed. The area of MUP or M-wave normalized in respect of the amplitude of the terminal phase (that is produced during extinction of the depolarized zones at the ends of the fibres) could be useful as a fatigue index. Analysing literature data on IAP changes due to Ca(2+) increasing, we hypothesised that the ability of muscle fibres to uptake Ca(2+) back into the sarcoplasmic reticulum could be the limiting site for fatigue. If this hypothesis is valid, IAP changes are not a cause of fatigue; they are due to it.     

Real-time kinetic modeling of voltage-gated ion channels using dynamic clamp

We propose what to our knowledge is a new technique for modeling the kinetics of voltage-gated ion channels in a functional context, in neurons or other excitable cells. The principle is to pharmacologically block the studied channel type, and to functionally replace it with dynamic clamp, on the basis of a computational model. Then, the parameters of the model are modified in real time (manually or automatically), with the objective of matching the dynamical behavior of the cell (e.g., action potential shape and spiking frequency), but also the transient and steady-state properties of the model (e.g., those derived from voltage-clamp recordings). Through this approach, one may find a model and parameter values that explain both the observed cellular dynamics and the biophysical properties of the channel. We extensively tested the method, focusing on Na_v models. Complex Markov models (10-12 states or more) could be accurately integrated in real time at >50 kHz using the transition probability matrix, but not the explicit Euler method. The practicality of the technique was tested with experiments in raphe pacemaker neurons. Through automated real-time fitting, a Hodgkin-Huxley model could be found that reproduced well the action potential shape and the spiking frequency. Adding a virtual axonal compartment with a high density of Na_v channels further improved the action potential shape. The computational procedure was implemented in the free QuB software, running under Microsoft Windows and featuring a friendly graphical user interface.     

Steady state analysis of influx and transbilayer distribution of ergosterol in the yeast plasma membrane

Mathematical modeling is now frequently used in outbreak investigations to understand underlying mechanisms of infectious disease dynamics, assess patterns in epidemiological data, and forecast the trajectory of epidemics. However, the successful application of mathematical models to guide public health interventions lies in the ability to reliably estimate model parameters and their corresponding uncertainty. Here, we present and illustrate a simple computational method for assessing parameter identifiability in compartmental epidemic models. We describe a parametric bootstrap approach to generate simulated data from dynamical systems to quantify parameter uncertainty and identifiability. We calculate confidence intervals and mean squared error of estimated parameter distributions to assess parameter identifiability. To demonstrate this approach, we begin with a low-complexity SEIR model and work through examples of increasingly more complex compartmental models that correspond with applications to pandemic influenza, Ebola, and Zika. Overall, parameter identifiability issues are more likely to arise with more complex models (based on number of equations/states and parameters). As the number of parameters being jointly estimated increases, the uncertainty surrounding estimated parameters tends to increase, on average, as well. We found that, in most cases, R0 is often robust to parameter identifiability issues affecting individual parameters in the model. Despite large confidence intervals and higher mean squared error of other individual model parameters, R0 can still be estimated with precision and accuracy. Because public health policies can be influenced by results of mathematical modeling studies, it is important to conduct parameter identifiability analyses prior to fitting the models to available data and to report parameter estimates with quantified uncertainty. The method described is helpful in these regards and enhances the essential toolkit for conducting model-based inferences using compartmental dynamic models.     

A Statistical Model for In Vivo Neuronal Dynamics

Single neuron models have a long tradition in computational neuroscience. Detailed biophysical models such as the Hodgkin-Huxley model as well as simplified neuron models such as the class of integrate-and-fire models relate the input current to the membrane potential of the neuron. Those types of models have been extensively fitted to in vitro data where the input current is controlled. Those models are however of little use when it comes to characterize intracellular in vivo recordings since the input to the neuron is not known. Here we propose a novel single neuron model that characterizes the statistical properties of in vivo recordings. More specifically, we propose a stochastic process where the subthreshold membrane potential follows a Gaussian process and the spike emission intensity depends nonlinearly on the membrane potential as well as the spiking history. We first show that the model has a rich dynamical repertoire since it can capture arbitrary subthreshold autocovariance functions, firing-rate adaptations as well as arbitrary shapes of the action potential. We then show that this model can be efficiently fitted to data without overfitting. We finally show that this model can be used to characterize and therefore precisely compare various intracellular in vivo recordings from different animals and experimental conditions.     

Analysis of the peak-to-peak ratio of extracellular potentials in the proximity of excitable fibres

In a previous work we studied the ratio between the amplitudes of the second and first phases (which we call PPR, after peak-to-peak ratio) of the single fibre action potential (SFAP) for a collection of fibrillation potentials (FPs) extracted from two pathological muscles. These FPs showed a wider PPR range than the Dimitrov–Dimitrova (D–D) convolutional model could provide. We proposed a modification of the D–D intracellular action potential (IAP) in order to obtain a range of PPRs comparable to that observed in our FPs. This paper extends that study to a large number of SFAPs extracted from the tibialis anterior muscle of normal subjects. The estimation of the average PPR range of non-diseased muscles in non-fatigued conditions is important since it can be used as a reference to establish a comparison with PPR ranges from muscles suffering some disorder or from fibres that are fatigued. Other aspects of the PPR, as its sensitivity with volume conductor parameters or to what extent changes in the SFAP PPR reflects changes in IAP spatial profile are also examined. We found that the PPR of experimental SFAPs ranges from 0.3 to 2.5 in all subjects and that all PPR histograms contain a well-defined single peak around the PPR value 1.0.     

A gradient model of cardiac pacemaker myocytes

We have formulated a spatial-gradient model of action potential heterogeneity within the rabbit sinoatrial node (SAN), based on cell-specific ionic models of electrical activity from its central and peripheral regions. The ionic models are derived from a generic cell model, incorporating five background and exchange currents, and seven time-dependent currents based on three- or four-state Markov schemes. State transition rates are given by non-linear sigmoid functions of membrane potential.

By appropriate selection of parameters, the generic model is able to accurately reproduce a wide range of action potential waveforms observed experimentally. Specifically, the model can fit recordings from central and peripheral regions of the SAN with RMS errors of 0.3987 and 0.7628 mV, respectively. Using a custom least squares parameter optimisation routine, we have constructed a spatially-varying gradient model that exhibits a smooth transition in action potential characteristics from the central to the peripheral region, whilst ensuring individual membrane currents remain physiologically accurate. Smooth transition action potential characteristics include maximum diastolic potential, overshoot potential, upstroke velocity, action potential duration and cycle length. The gradient model is suitable for developing higher dimensional models of the right atrium, in which action potential heterogeneity within nodal tissue may be readily incorporated.     

Analysis of the relationship between the rise-time and the amplitude of single-fibre potentials in human muscles

Using the core-conductor theory, a single fibre action potential (SFAP) can be expressed as the convolution of a biolectrical source and a weight function. In the Dimitrov–Dimitrova (D–D) SFAP convolutional model, the first temporal derivative of the intracellular action potential (IAP) is used as the source. The present work evaluates the relationship between the SFAP peak-to-peak amplitude (V_pp) and peak-to-peak interval (rise-time, RT) at different fibre-to-electrode distances using simulated signals obtained by the D–D model as well as real recordings. With a single fibre electrode, we recorded 63 sets of consecutive SFAPs from the m. tibialis anterior of four normal subjects. The needle was intentionally moved whilst recording each SFAP set. We used the observed changes in RT and V_pp within each SFAP set as a point of reference with which to evaluate how closely the relationship between RT and V_pp provided by the D–D model reflects real data. We found that half of the recorded SFAP sets had rise-times higher than those generated by the D–D model. We also showed the influence of the IAP spatial length on the sensitivity of RT and V_pp with radial distance. The study reveals some inaccuracies in simulated SFAPs whose origin might be related to the assumptions made in the core-conductor theory.     

Initiation of Sodium Spikelets in Basal Dendrites of Neocortical Pyramidal Neurons

Cortical information processing relies critically on the processing of electrical signals in pyramidal neurons. Electrical transients mainly arise when excitatory synaptic inputs impinge upon distal dendritic regions. To study the dendritic aspect of synaptic integration one must record electrical signals in distal dendrites. Since thin dendritic branches, such as oblique and basal dendrites, do not support routine glass electrode measurements, we turned our effort towards voltage-sensitive dye recordings. Using the optical imaging approach we found and reported previously that basal dendrites of neocortical pyramidal neurons show an elaborate repertoire of electrical signals, including backpropagating action potentials and glutamate-evoked plateau potentials. Here we report a novel form of electrical signal, qualitatively and quantitatively different from backpropagating action potentials and dendritic plateau potentials. Strong glutamatergic stimulation of an individual basal dendrite is capable of triggering a fast spike, which precedes the dendritic plateau potential. The amplitude of the fast initial spikelet was actually smaller that the amplitude of the backpropagating action potential in the same dendritic segment. Therefore, the fast initial spike was dubbed “spikelet”. Both the basal spikelet and plateau potential propagate decrementally towards the cell body, where they are reflected in the somatic whole-cell recordings. The low incidence of basal spikelets in the somatic intracellular recordings and the impact of basal spikelets on soma-axon action potential initiation are discussed.     

Effects of changes in the shape of the intracellular action potential on the peak-to-peak ratio of single muscle fibre potentials

In situ recording of the intracellular action potential (IAP) of human muscle fibres is not yet feasible, and consequently, knowledge about certain IAP characteristics of these IAPs is still limited. The ratio between the amplitudes of the second and first phases (the so-called peak-to-peak ratio, PPR) of a single fibre action potential (SFAP) is known to be closely related to the IAP profile. The PPR of experimentally recorded SFAPs has been found to be largely independent of changes in the fibre-to-electrode (radial) distance. The main goal of this paper is to analyze the effect of changes in different aspects of the IAP spike on the relationship between PPR and radial distance. Based on this analysis, we hypothesize about the characteristics of IAPs obtained experimentally. It was found that the sensitivity of the SFAP PPR to changes in radial distance is essentially governed by the duration of the IAP spike. Assuming that, for mammals, the duration of the IAP rising phase lies within the range 0.2-0.4 ms, we tentatively suggest that the duration of the IAP spike should be over approximately 0.75 ms, with the shape of the spike strongly asymmetric. These IAP characteristics broadly coincide with those observed in mammal IAPs.     

Estimation of population firing rates and current source densities from laminar electrode recordings

This model study investigates the validity of methods used to interpret linear (laminar) multielectrode recordings. In computer experiments extracellular potentials from a synaptically activated population of about 1,000 pyramidal neurons are calculated using biologically realistic compartmental neuron models combined with electrostatic forward modeling. The somas of the pyramidal neurons are located in a 0.4 mm high and wide columnar cylinder, mimicking a stimulus-evoked layer-5 population in a neocortical column. Current-source density (CSD) analysis of the low-frequency part (<500 Hz) of the calculated potentials (local field potentials, LFP) based on the ‘inverse’ CSD method is, in contrast to the ‘standard’ CSD method, seen to give excellent estimates of the true underlying CSD. The high-frequency part (>750 Hz) of the potentials (multi-unit activity, MUA) is found to scale approximately as the population firing rate to the power 3/4 and to give excellent estimates of the underlying population firing rate for trial-averaged data. The MUA signal is found to decay much more sharply outside the columnar populations than the LFP.     

Quantifying spatial and temporal variation of North Pacific fin whale (Balaenoptera physalus) acoustic behavior

In order to help develop hypotheses of connectivity among North Pacific fin whales, we examine recordings from 10 regions collected in the spring and fall. We develop a Random Forest model to classify fin whale note types that avoids manual note classification errors. We also present a method that objectively quantifies the note and pattern composition of recordings. We find that fin whale recordings near Hawaii have distinctive patterns, similar to those found in other regions in the central North Pacific, suggesting potential migration pathways. Our results are consistent with previous studies that suggest there may be two different populations utilizing the Chukchi Sea and central Aleutians in the fall and mix to some degree in the southern Bering Sea. Conversely, we found little difference between spring and fall recordings in the eastern Gulf of Alaska, suggesting some residency of whales in this region. This is likely due to fine scale similarities of calls among the inshore regions of British Columbia, while offshore areas are being utilized by whales traveling from various distant areas. This study shows how our novel approach to characterize recordings is an objective and informative way to standardize spatial and temporal comparisons of fin whale recordings.     

A Comparative Survey of Automated Parameter-Search Methods for Compartmental Neural Models

One of the most difficult and time-consuming aspects of building compartmental models of single neurons is assigning values to free parameters to make models match experimental data. Automated parameter-search methods potentially represent a more rapid and less labor-intensive alternative to choosing parameters manually. Here we compare the performance of four different parameter-search methods on several single-neuron models. The methods compared are conjugate-gradient descent, genetic algorithms, simulated annealing, and stochastic search. Each method has been tested on five different neuronal models ranging from simple models with between 3 and 15 parameters to a realistic pyramidal cell model with 23 parameters. The results demonstrate that genetic algorithms and simulated annealing are generally the most effective methods. Simulated annealing was overwhelmingly the most effective method for simple models with small numbers of parameters, but the genetic algorithm method was equally effective for more complex models with larger numbers of parameters. The discussion considers possible explanations for these results and makes several specific recommendations for the use of parameter searches on neuronal models.     

Activity-mediated accumulation of potassium induces a switch in firing pattern and neuronal excitability type

During normal neuronal activity, ionic concentration gradients across a neuron’s membrane are often assumed to be stable. Prolonged spiking activity, however, can reduce transmembrane gradients and affect voltage dynamics. Based on mathematical modeling, we investigated the impact of neuronal activity on ionic concentrations and, consequently, the dynamics of action potential generation. We find that intense spiking activity on the order of a second suffices to induce changes in ionic reversal potentials and to consistently induce a switch from a regular to an intermittent firing mode. This transition is caused by a qualitative alteration in the system’s voltage dynamics, mathematically corresponding to a co-dimension-two bifurcation from a saddle-node on invariant cycle (SNIC) to a homoclinic orbit bifurcation (HOM). Our electrophysiological recordings in mouse cortical pyramidal neurons confirm the changes in action potential dynamics predicted by the models: (i) activity-dependent increases in intracellular sodium concentration directly reduce action potential amplitudes, an effect typically attributed solely to sodium channel inactivation; (ii) extracellular potassium accumulation switches action potential generation from tonic firing to intermittently interrupted output. Thus, individual neurons may respond very differently to the same input stimuli, depending on their recent patterns of activity and/or the current brain-state.     

Action potentials of the rabbit, guinea pig, dog and albino rat working ventricular myocardium under interpolated extrasystole conditions during postnatal ontogenesis

Experiments were carried out on the working myocardium of the right heart ventricle of newborn and adult rabbits, guinea-pigs, dogs and albino rats. In the dog, the guinea-pig and the rabbit, after ten action potentials (AP) elicited with 1 Hz frequency we always interpolated an extrasystole at an interval (TE) of 100-900 ms. In albino rats we used a basic frequency of 2 Hz and a TE of 30-370 ms from the last regular AP. Using glass microelectrodes, we recorded the extrasystolic AP (EAP) and the next subsequent AP (2AP). The results were evaluated by constructing graphs of the correlations of the duration of the plateau phase (D0) to TE and of the duration of repolarization to -60 mV level (D60) to the TE. In the myocardium of newborn rabbits, guinea-pigs and dogs, with short TE both D0 and D60 of the EAP are shorter than in the steady state (SS), while for the 2AP the same parameters are influenced only a little. As the TE lengthens, the EAP gradually acquire a length corresponding more to the SS. With TE longer than half the duration of the cycle in the steady state the EAP return to normal, while the 2AP become shorter. The effect of extrasystole on the rat EAP and 2AP diminished with advancing age. In the myocardium of adult rabbits and adult guinea-pigs, and slightly in the myocardium of adult dogs and newborn rats, we observed that the duration of the EAP, with certain TE, was greater than in the steady state.(ABSTRACT TRUNCATED AT 250 WORDS)     

Dynamical estimation of neuron and network properties II: path integral Monte Carlo methods

Hodgkin–Huxley (HH) models of neuronal membrane dynamics consist of a set of nonlinear differential equations that describe the time-varying conductance of various ion channels. Using observations of voltage alone we show how to estimate the unknown parameters and unobserved state variables of an HH model in the expected circumstance that the measurements are noisy, the model has errors, and the state of the neuron is not known when observations commence. The joint probability distribution of the observed membrane voltage and the unobserved state variables and parameters of these models is a path integral through the model state space. The solution to this integral allows estimation of the parameters and thus a characterization of many biological properties of interest, including channel complement and density, that give rise to a neuron’s electrophysiological behavior. This paper describes a method for directly evaluating the path integral using a Monte Carlo numerical approach. This provides estimates not only of the expected values of model parameters but also of their posterior uncertainty. Using test data simulated from neuronal models comprising several common channels, we show that short (<50 ms) intracellular recordings from neurons stimulated with a complex time-varying current yield accurate and precise estimates of the model parameters as well as accurate predictions of the future behavior of the neuron. We also show that this method is robust to errors in model specification, supporting model development for biological preparations in which the channel expression and other biophysical properties of the neurons are not fully known.     

Capillaries within compartments: microvascular interpretation of dynamic positron emission tomography data

Measurement of exchange of substances between blood and tissue has been a long-lasting challenge to physiologists, and considerable theoretical and experimental accomplishments were achieved before the development of the positron emission tomography (PET). Today, when modeling data from modern PET scanners, little use is made of earlier microvascular research in the compartmental models, which have become the standard model by which the vast majority of dynamic PET data are analysed. However, modern PET scanners provide data with a sufficient temporal resolution and good counting statistics to allow estimation of parameters in models with more physiological realism. We explore the standard compartmental model and find that incorporation of blood flow leads to paradoxes, such as kinetic rate constants being time-dependent, and tracers being cleared from a capillary faster than they can be supplied by blood flow. The inability of the standard model to incorporate blood flow consequently raises a need for models that include more physiology, and we develop microvascular models which remove the inconsistencies. The microvascular models can be regarded as a revision of the input function. Whereas the standard model uses the organ inlet concentration as the concentration throughout the vascular compartment, we consider models that make use of spatial averaging of the concentrations in the capillary volume, which is what the PET scanner actually registers. The microvascular models are developed for both single- and multi-capillary systems and include effects of non-exchanging vessels. They are suitable for analysing dynamic PET data from any capillary bed using either intravascular or diffusible tracers, in terms of physiological parameters which include regional blood flow.     

Stochastic Dynamic Model for Estimation of Rate Constants and Their Variances from Noisy and Heterogeneous PET Measurements

Tissue heterogeneity, radioactive decay and measurement noise are the main error sources in compartmental modeling used to estimate the physiologic rate constants of various radiopharmaceuticals from a dynamic PET study. We introduce a new approach to this problem by modeling the tissue heterogeneity with random rate constants in compartment models. In addition, the Poisson nature of the radioactive decay is included as a Poisson random variable in the measurement equations. The estimation problem will be carried out using the maximum likelihood estimation. With this approach, we do not only get accurate mean estimates for the rate constants, but also estimates for tissue heterogeneity within the region of interest and other possibly unknown model parameters, e.g. instrument noise variance, as well. We also avoid the problem of the optimal weighting of the data related to the conventionally used weighted least-squares method. The new approach was tested with simulated time–activity curves from the conventional three compartment – three rate constants model with normally distributed rate constants and with a noise mixture of Poisson and normally distributed random variables. Our simulation results showed that this new model gave accurate estimates for the mean of the rate constants, the measurement noise parameter and also for the tissue heterogeneity, i.e. for the variance of the rate constants within the region of interest.