HARTLINE LAB SELECTED PUBLICATION SUMMARIES
IN COMPUTATIONAL NEUROSCIENCE
IN COPEPOD NEUROECOLOGY
- Lenz, Hartline and Davis (2000) (J Comp. Physiol. A)
- Weatherby, Davis, Hartline and Lenz (2000) (J Comp. Physiol. A)
- Hartline, Buskey and Lenz (1999) (Biol. Bull.)
- Davis, Weatherby, Hartline and Lenz (1999) (Nature)
- Lenz and Hartline (1999) (Mar. Biol.)
- Hartline, Lenz and Herren (1996) (MFBP)
- Gassie, Lenz, Yen and Hartline (1993)
- Yen, Lenz, Gassie and Hartline (1992)
"Simulations of voltage clamping poorly space-clamped voltage-dependent conductances
in a uniform cylindrical neurite"
Hartline, D.K. and Castelfranco, A.M.
J. comput. Neuroscience (2003)
Significant error is made by using a point voltage clamp to measure active ionic current
properties in poorly space-clamped cells. This can even occur when there are no obvious
signs of poor spatial control. We evaluated this error for experiments that employ an
isochronal I(V) approach to analyzing clamp currents. Simulated voltage clamp experiments
were run on a model neuron having a uniform distribution of a single voltage-gated
inactivating ionic current channel along an elongate, but electrotonically compact, process.
Isochronal Boltzmann I(V) and kinetic parameter values obtained by fitting the Hodgkin-Huxley
equations to the clamp currents were compared with the values originally set in the model.
Good fits were obtained for both inward and outward currents for moderate channel densities.
Most parameter errors increased with conductance density. The activation rate parameters
were more sensitive to poor space clamp than the I(V) parameters. Large errors can occur
despite "normal"- looking clamp curves.
"Simulations of space-clamp errors in estimating parameters of voltage-gated conductances
localized at different electrotonic distances"
A.M. Castelfranco and D.K. Hartline
Neurocomputing (2002: in press)
Abstract: Parameters of voltage-gated conductances measured in point voltage clamp
experiments are in error if the membrane containing the conductance is under poor spatial control.
The ability to correct such errors could rescue much flawed experimental data, but requires a
detailed understanding of the errors. We evaluated such errors using simulated voltage-clamp
experiments on a soma with a single (0.5 lambda) cylindrical process having a patch of
voltage-dependent Hodgkin-Huxley channels placed various distances from the soma.
Kinetic and steady-state parameters were obtained by curve-fitting the kinetics of soma clamp
current flowing in response to step command depolarizations. Most errors in fitted parameters
increased as the patch was located more distally on the neurite.
Multistable neurons: ionic mechanisms of bursting
Daniel K. Hartline
Sensory Systems of Arthropods eds Wiese, K., Gribakin, F., Popov, A.V.
and Birkhauser, G.H. 557-566 (1993)
Increasingly, neurons are being found that have two quasi-stable states,
high-frequency firing or silence, either of which may be present only transiently.
Bursts of spikes can be triggered or terminated spontaneously, or by brief inputs,
giving the output of the neuron some of the properties of an electronic monostable
multivibrator. A voltage-dependent inward current giving rise to a regenerative
"plateau potential" underlies this behavior in many neurons, arthropod and
non-arthropod. Inactivation of inward current or slow activation of outward
current may spontaneously terminate plateaus. Control of plateaus often involves
a modulatory transmitter coupled to second messenger production. This mechanism can
"reprogram" a neuron from a quasi-linear one into a bursty, bistable one.
Effects of soma isolation on outward currents measured under voltage clamp
in spiny lobster stomatogastric motor neurons
Daniel K. Hartline, Donald V. Gassie, and Bradley R. Jones
Journal of Neurophysiology 69(6): 2056-2071 (1993)
1. Outward currents in identified cell types from the pyloric system of the
stomatogastric ganglion (STG) of the spiny lobster,Panulirus marginatus,
were studied under two-microelectrode voltage clamp. A comparison was made between
data from intact cells and somata isolated by ligation of the primary neurite of
these monopolar neurons.
2. Despite the elimination of current contributions from the extensive arborizations
of STG neurons, few significant differences were found in the mean values of
parameters for outward currents between populations of isolated somata and intact
cells of a given type. Measurements that showed little difference included magnitude
and activation threshold of a calcium-dependent outward current (IJ) and magnitude,
activation threshold, voltage dependence, and inactivation time course of A current
(IA). Although previous work has suggested that IJ, might reside predominantly in
the soma, IA is known to be distributed in poorly space-clamped neurite processes.
The absence of obvious effects of isolation was thus unexpected.
3. To better understand the mechanisms involved, we used compartmental models derived
from reconstructed neurons to simulate the effects of isolation. It was concluded that,
for the particular conditions present in stomatogastric neurons, with a large, uniformly
distributed outward current conductance activated, even though neurites and axon remain
attached, most measured current flows through well-clamped soma membrane.
4. Factors contributing to this result included the outward sign of the current, the
large specific conductance activated in these neurons (among the larger reported in somata),
and the presence of only a single major process leaving the soma. The potential for
serious errors in voltage-clamp measurements from intact cells remains if these
conditions are not met.
Simulation of peptide processing, compartmentation and release in neurosecretory cells
Daniel K. Hartline and Robert W. Newcomb
Neurochem. Int. 19(3): 281-296 (1991)
A theoretical framework is presented which mathematically describes the transport of secretory
granules in neurons and the enzymatic processing of a prohormone and subsequent intermediates within
such granules. The transport system represents the synthesis of individual granules, the migration of
the granules among various ''diffusionally"-connected compartments, and the release of granules from
designated compartmental release sites. The processing of prohormone is represented by a multiple-site
cleavage reaction with separate rate constants for each cleavage site on each peptide fragment. The effects
of neuron growth, processing enzyme activity decay, and changes in release probability with granule age
are taken into account. Computer models were constructed based on these underlying assumptions.
Simulations are used to illustrate how the distribution or peptide fragments contained in a neuron will
depend on interactions of peptide processing kinetics and cellular transport and storage processes. The
models provide a tool for testing how multiple cell-biological variables will interact to control the chemical
mixture secreted from a peptidergic neuron.
Quantitative analysis and computer simulation of oxytocin-neurophysin processing in the rat neurohypophysis
Robert W. Newcomb, Daniel K. Hartline, Jean-Georges Lorentz, Antoine Depaulis and Jean J. Nordmann
Neurochem. Int. 19(3): 297-312 (1991)
A chromatographic procedure is used to obtain data on amounts of the oxytocin-neurophysin
(OT-NP, or "A") and its more completely processed form (OT-NP', or "B"), which lacks the carboxyl
terminal glutamic acid residue of the A form. Measurements of B +A and B/(B + A) are made for single
neurohypophyses under varying physiological conditions, in release obtained by perfusion of isolated
neurohypophysis with saline containing increased external potassium concentrations, and in isolated neural
lobe nerve endings and nerve swellings. The chromatographic procedure is also combined with injection
of [35S]cysteine over the supraoptic nucleus. This approach is used to obtain the in vivo A to B
conversion rate, and to investigate the trafficking of newly synthesized secretory granules in
different axonal compartments. The manner with which the conversion rate changes with secretory
activity is investigated by comparing the conversion rate in control animals to that in animals
subjected to osmotic stress. Generalized computer models of peptide processing, storage,
and release in neurons are used to test various hypotheses of the control of secretory
glanule movements in the neurohypophysis for sufficiency in explaining the changes in
B + A and B/(B +A) with changing physiological state.
Voltage clamp analysis of intact stomatogastric neurons
Katherine Graubard and Daniel K. Hartline
Brain Research 557: 241-254 (1991)
Two-electrode voltage clamp of intact, identified pyloric neurons of the spiny lobster
stomatogastric ganglion reveals two major outward currents. A rapidly inactivating,
tetraethylammonium- (TEA) insensitive, 4-aminopyridine- (4AP) sensitive, outward current
resembles IA of molluscan neurons; it activates rapidly on depolarizations above rest
(e.g. -45 mV), delaying both thc axonal-sodium and the neuropil calcium spikes which
escape voltage-clamp control. We infer that A-current is distributed both in a space
clamped region (on or near the soma) and in a non-space clamped region with access
to the generators for sodium and calcium spikes. A calcium-dependent outward current,
Io(Ca), activates rapidly at clamp steps above -25 mV and inactivates at depolarizing
holding voltages. Increasing depolarization results in an increase in both Io(Ca) and
firing rate but a reduction in the amplitude of the sodium spike current. Blockage of Io(Ca)
with Cd++ causes little change in spike firing pattern. These observations are consistent
with Io(Ca) being activated primarily in the soma and nearby regions which are under
good control with a soma voltage clamp (and distant from the Na+-spike trigger zone).
While the lack of space clamp limits resolution of charging transients and tail currents,
the identification of the major current subgroups can still be readily accomplished,
and inferences about the location and function of currents can be made which would
not be possible if the cells were space clamped or truncated.
Simulation of restricted neural networks with reprogrmmable neurons
Daniel K. Hartline
IEEE Transactions on circuits and systems 36:(5) 653-660 (1989)
This paper describes a network model composed of reprogrammable
neurons, incorporating the following design features:
- Spikes can be generated by a model representing repetitive firing at
axon (and dendritic) trigger zones.
- Active responses (plateau potentials; delaying mechanisms) are simulated
with Hodgkin-Huxley type kinetics.
- Synaptic interactions, both spike-mediated and non-spiking chemical
("chemotonic"), simulate transmitter release and binding to postsynaptic
receptors. Facilitation and antifacilitation of spike-mediated postsynaptic
potentials (PSP's) are included.
- Chemical pools are used to simulate second messenger systems,
trapping of ions in extracelluar spaces, and electrogenic pumps, as well
as biochemical reaction chains of quite general character. Modulation of any
of the parameters of any compartment can be effected through the pools.
- Intracellular messengers of three kinds are simulated explicitly: a) those
produced by voltage-gated processes (e. g., Ca ); b) those dependent on transmitter
(or hormone) binding; and c) those dependent on other internal messengers (e. g.,
internally released Ca; enzymatically activated pathways).
Full-wave rectification from a mixed electrical chemical synapse
Katherine Graubard and Daniel K. Hartline
Science 237: 535-537(1987)
Electrical and chemical synapses usually reinforce one another, but the pyloric late-to-
lateral pyloric (PL-to-LP) neuronal connections in lobster stomatogastric ganglia
create an inverted U-shaped transfer function between the two neurons: regardless of
whether the PL membrane voltage swings positive or negative, the postsynaptic LP
voltage will go negative. When the presynaptic cell voltage goes negative, the effect on
the LP voltage is due to electrical coupling. During positive presynaptic voltages, the
strong contribution of graded chemical inhibition from the PL to the LP neuron
overrides the positive electrical coupling to produce net negativity.
A microprocessor-controlled stimulator for generating voltage clamp command sequences
Chien Chang, Ning Hsu and Daniel K. Hartline
This paper presents a design for a programmable stimulator which delivers a
systematically varying sequence of command pulses to a voltage clamp or "current
clamp" apparatus. The need for such stimulus sequencing is typical of many
applications in neurophysiology. In voltage clamp studies, for example, the effects
of "conditioning" pulses of varying magnitudes or durations will be examined on
responses to subsequent "test" pulses. Such paradigms are used to characterize the
activation and inactivation characteristics of ionic mechanisms. Although a labora-
tory computer is easily programmed for such applications, its disadvantages in cost,
size, ease of use, and competing demands make alternatives attractive. The micro-
processor-controlled stimulator presented here has much of the flexibility of
a laboratory computer in stimulus control, but it is compact, portable, convenient to
use and economical.
Journal of Neuroscience Methods 18: 361-370 (1986)
Synaptic regulation of cellular properties and burst oscillations of neurons in
gastric mill system of spiny lobsters, Panulirus interruptus
David F. Russell and Daniel K. Hartline
Journal of Neurophysiology 52(1): 54-72 (1984)
1. The properties of neurons in the stomatogastric ganglion (STG)
participating in the pattern generator for the gastric mill rhythm were
studied by intracellular current injection under several conditions:
during on-going gastric rhythms, in the nonrhythmic isolated STG, after
stimulation of the nerve carrying central nervous system (CNS) inputs
to the STG, or under Ba2++ or Sr2++.
2. Slow regenerative depolarizations during ongoing rhythms were
demonstrated in the interior median, cardiopyloric, lateral cardiac,
gastropyloric, and continuous inhibitor (AM, CP, LC, GP, and Cl) neurons
according to criteria such as voltage dependency, burst triggering,
and termination by brief current pulses, etc. Experiments showed that
regenerative like behavior was not due to synaptic network interactions.
3. The slow regenerative responses were abolished by isolating the
stomatogastric ganglion but could be reestablished by stimulating
the input nerve. This indicates that certain CNS inputs synaptically
induce the regenerative property in specific gastric neurons.
4. Slow regenerative depolarizations were not demonstrable in gastric mill (GM)
motor neurons. Their burst oscillations and firing
rate were instead proportional to injected current. CNS inputs evoked a prolonged
depolarization in GM motor neurons, apparently by a nonregenerative mechanism.
5. All the gastric cells showed prolonged regenerative potentials under
0.5 -1.5 mM Ba2++.
6. We conclude that the gastric neurons of the STG can be divided into three
types according to their properties: Those with a regenerative
capability, a repetitively firing type, and a nonregenerative "proportional"
type. The cells are strongly influenced by several types of CNS inputs,
including "gastric command fibers."
Endogenous burst capability in a neuron of the gastric mill pattern generator
of the spiny lobster Panulirus interruptus
Daniel K. Hartline and David F. Russell
Journal of Neurobiology 15(5): 345-364(1984)
The gastric system of the lobster stomatogastric ganglion has previously
been thought to include no neurons capable of endogenous bursting. We
describe conditions under which one of the motorneurons, the CP cell, can
burst endogenously in a free-running manner in the absence of other phasic
network activity. Isolated preparations of the foregut nervous system were
used, and the CP bursting was either spontaneous or was activated by
continuous stimulation of an input nerve.
Three criteria were applied to establish the endogenous nature of such
burst generation in CP: absence of phasic input, reset of the bursting pattern
by pulses of current in a characteristic phase-dependent manner, and modulation
of burst rate by sustained injected current. (l) The firing of other
cells which are known to be related synaptically to CP was monitored in
nerve records. These other cells were either silent or fired only tonically.
Cross-correlograms showed that CP bursting was not ascribable to phasic
activity in these other network cells. (2) A depolarizing current pulse of
sufficient strength injected intracellularly between bursts triggered a burst
prematurely and reset the subsequent rhythm. A hyperpolarizing pulse
during a burst terminated it and reset the subsequent rhythm. Reset behavior
was similar to that described for other endogenous bursters. (3) Application
of a positive-going ramp current initially caused an increase in burst rate,
as described for other endogenous bursters. However, further depolarization
caused a slower burst rate due to lengthening of the individual bursts,
although mean firing frequency continued to increase throughout the range
Such free-running endogenous repetitive bursting appeared to result from
the CP's ability to produce slow regenerative depolarizations ("plateau
potentials"). When bursting was present, so was the plateau property, as
determined by I-V analysis and by the ability of brief current pulses to trigger
and terminate bursts. The previous inability to observe endogenous bursting
in preparations with central input removed may be due to the usual absence
of the plateau property in such preparations.
CP bursting during normal gastric rhythms, while underlain by plateau potentials,
is strongly controlled by network interactions. CP appears not to be well placed
in the network to be considered a source of normal gastric rhythmicity. Nevertheless,
endogenous bursting in CP may explain some of the partial gastric rhythms seen in
behavioral studies, and illustrates one way that cellular properties might contribut
to rhythmic behaviors.
Graded synaptic transmission between identified spiking neurons
Katherine Graubard, Jonathan A. Raper, and Daniel K. Hartline
Journal of Neurophysiology 50(2): 508-521(1983)
1. Graded synaptic transmission between spiking
motoneurons of the pyloric group was studied in
the stomatogastric ganglion of the spiny lobster, Panulirus
interruptus. Intracellular microelectrodes were placed in
the cell bodies of both pre- and postsynaptic neurons.
2. Graded synaptic transmission was found between all
tested cell pairs that were known to display spike-evoked
synaptic transmission, including PD to LP, PD to PE,
PD to PL, PL to LP, and LP to PD.
3. Graded synaptic transmission was effective below the
threshold for spikes. Thus, it was possible to study the
influence of graded synaptic transmission in normally
active ganglia without blockage of spikes by tetrodotoxin.
PD and LP neurons that were known to produce spike-evoked
inhibitory postsynaptic potentials (IPSPs) were also
capable of producing inhibitory effects on postsynaptic
cells below the threshold for spikes.
4. When tetrodotoxin (TTX) was used to eliminate both
spikes and endogenous membrane oscillations, depolanzation
of presynaptic neurons produeed hyperpolarization of postsynaptic
cells. The presynaptic response to a current step usually showed
a small early peak and a maintained, slightly lower plateau. The
postsynaptic response had a delay,then a rise to a pronounced peak,
and a roughly exponential decline to a maintained plateau. There
was a presynaptic voltage threshold for any postsynaptic response;
beyond the threshold, both pre- and postsynaptic peak and plateau
responses increased with increasing current. PD neurons normally
are depolarized beyond their release threshold in tetrodotoxin and,
thus, released transmitter tonically for the many-hour duration of
5. Chemical, tonic synaptic tranmission, here called graded synaptic
transmission, was demonstrated by the presence of the following
criteria: 1) reversal in sign of the postsynaptic response, 2) synaptic
delay, 3) reversal potential, 4) postsynaptic conductance increase,
5) graded and reversible block by reduction of external Ca++, and
6)specific graded block of the LP-to-PD synapse without effect on
the PD-to-LP synapse by less than 10 uM picrotoxin added to the bathing