Department of Biology, Wesleyan College, Macon, GA 31210.

1
Mutual Excitation Among Olfactory Bulb Mitral
Cells Revealed by Recurrence Time History
Matching (RTHM)
Alexandra Radu, Maame Boateng, Henaa Razzak,
Barry K. Rhoades.
Department of Biology, Wesleyan College, Macon,
GA 31210.
BACKGROUND Olfactory Bulb Function
METHODS Surgery and Monitoring
BACKGROUND Correlational Analysis
RESULTS Single Units
CONCLUSIONS
.
1) Interspike Interval Histograms provide weak
evidence for excitatory input to MCL cells. 2)
RTHM plots of simultaneously recorded
spiking cells from the MCL show evidence for
mutual synaptic excitation. 3) RTHM plots do not
provide any evidence for lateral inhibition
between mitral cells.
In the mammalian olfactory system, the primary
cortical sensory area is the olfactory bulb (OB).
An interspike interval histogram (ISIH) shows the
cumulative distribution of intervals between
action potentials in a unit spike train.
Surgical Methods Subjects Male Sprague-Dawley
rats (Rattus norvegicus) from 180 to
450g Anesthesia Anesthesia for each rat was
maintained at a level characterized by dilated
pupils, regular breathing, insensitivity to deep
leg pinch pain, and no movement, following one of
three anesthesia protocols 1) Urethane
(125mg/ml in normal saline) IP _at_ 1200-1500 mg/kg
body weight 2) Nembutal (sodium pentobarbital
50mg/ml) IP _at_ 50-75mg/kg body weight 3)
Ketamine cocktail (Ketamine 60mg/ml, Xylazine
6mg/ml, Acepromazine 1mg/ml) IM _at_
0.2-0.4ml plus Nembutal _at_ 0.2-0.3ml as required.
Surgical Exposure The following steps
were performed in sequence 1) the rat was
placed on a warming pad (Gaymar T/Pump), on a
mechanical isolation table within a
grounded mesh Faraday cage 2) the head was
secured in the nose clip and ear bars of
stereotaxic frame (David Kopf Instr.) 3) the
scalp, superficial muscle, and fascia were cut
down the midline and reflected laterally
to expose the dorsal skull 4) two stainless
steel screws for electrical ground and reference
were implanted in the parietal and/or
nasal bones 5) the left orbit was surgically
exenterated of all soft tissue 6) a surgical
drill was used to open a window encompassing the
lateral OB, lateral olfactory tract,
and anterior pole of the frontal cortex 7) the
dura mater was stripped off and the exposed brain
surface was coated with mineral
oil Electrophysiological Methods Electrode
Placement Two electrodes were stereotaxically
positioned 1) a 250um formvar-insulated
stainless steel wire loop was placed against the
OB surface to serve as a surface EEG
electrode and presser foot to suppress vascular
pulsations 2) a electrolytically-sharpened,
parlene-insulated tungsten-steel microelectrode
(A-M Systems 575500, 5Mohm) was lowered
to depth through the center of the presser-foot
ring Amplification and Monitoring The signal
from the depth microelectrode was passed through
a microelectrode amplifier (A-M Systems 1800
x1000 500Hz-5kHz band-pass filtered) and
simultaneously monitored in three ways 1)
computer video display via an A/D conversion
system (ADInstruments PowerLab 8sp Scope
5.11 software) 2) audio monitoring (RCA model
SA-155 stereo amplifier) 3) analog
oscilloscope (GW model GOS-622G)
Representative Spike Trains From Two Rats
TGase1 Positive
rat brain
Primary olfactory nerve (PON) axons synapse on
mitral (M) and tufted (T) projection neurons.
Periglomerular (P) and granule (G) neurons
provide lateral interactions.
time (5 msec ticks)
ABSTRACT
Five-second epochs of single unit activity show
non-bursting firing.
Single Unit ISIHs Provide Weak Evidence for
Excitatory Input to MCL Cells
In the mammalian olfactory bulb coordinated
neural activity associated with odor
discrimination is dominated by narrow-band
oscillations in the gamma EEG range (30-80 Hz),
gated by respiratory inspiration. Mutual
excitation among the mitral cell projection
neurons is a required feature of successful
bulbar models, but has not been conclusively
demonstrated by either direct histological or
electrophysiological evidence. In the present
study tungsten-steel microelectrodes were lowered
to the mitral cell layer of the olfactory bulb in
urethane- or pentobarbital-anesthetized rats, and
positioned using electrophysiological response
criteria. Ten to twenty minute samples of
resting neuronal activity were recorded to analog
tape. A template matching system was used to
isolate and extract multiple, simultaneous
single-unit spike trains. Temporal linkages
between spike trains were evaluated using
recurrence-time history matching (RTHM), a
algorithm for enhancing conventional conditional
cross-correlograms. In the RTHM analysis plots
of some pairs of mitral cells apparent mutual
excitation was evidenced by the presence of
high-density, short-latency linkage bands and a
within-band pattern suggesting activity-dependent
gain. The validity of inferences based on RTHM
was verified using artificially-generated spike
trains with stochastically-defined firing
patterns and inter-neuronal linkages.
Two Theories of OB Function
a
b
11-1
A cross-correlogram is a histogram of cross
intervals between two unit spike trains recorded
concurrently. Intervals corresponding to
excitatory synaptic delays appear as linkage
peaks above the background correlational noise
level.
Shepherd (1979) Lateral negative feedback
sharpens chemotopic image of signal from
olfactory receptors. Evidence - PG cells are
GABAergic and GABA is an inhibitory
neurotransmitter - paired-pulse attenuation of
evoked potentials to strong PON stimulation
c
d
9-1
Cross Intervals
a, c Interspike interval histograms from mitral
cells approximate Poisson distributions truncated
by an initial refractory period. b, d A
stochastic model was used to generate ISIHs for
artificial spike trains (light traces) with
identical mean firing rates and refractory
periods to the original real spike trains (bold
traces). The real spike trains show a higher
number of very short intervals, as would be
expected for units participating in a mutually
excitatory neuronal ensemble.
Freeman (1975) MltgtG negative feedback creates
oscillatory activity ( 60Hz for rat). PltgtP and
MltgtM mutual lateral excitation promotes spatial
amplitude patterns with information
content. Evidence - averaged evoked potentials
(AEP) indicate PltgtP excitation - current
source density analysis indicates GL
excitation - PltgtP excitation mediated by GABA
- conventional correlograms show MltgtM
excitation (de la Luz, 2005)
time (10 msec ticks)
METHODS Data Criteria
Cross Correlogram
linkage peak
Anesthetic Level At an appropriate anesthetic
level, the OB surface EEG was characterized by
slow (1-3HZ) waves synchronized to respiratory
inspiration, with sinusoidal gamma frequency
oscillations (arrows) riding the recovery phase
of most waves.
.
RESULTS RTHM Plots
Experimental Question
MCL Cells Evidence Weak Mutual Excitation
BACKGROUND RTHM Analysis
Will enhanced spike train correlational analysis
of paired M cell single unit spike trains (via
RTHM) support Mltgt M mutual excitation, as
evidenced by linkage bands and activity-dependent
gain?
a
b
Identification of Mitral Cells for Recording As
the microelectrode was lowered into the OB, the
glomerular layer (GL) was identified by
multicellular burst firing (arrows) ,
synchronized to the respiratory wave. The
underlying mitral cell layer (MCL) was
characterized by isolated unit APs, with low
background activity, and non-synchronized,
non-bursting firing.
Recurrence time history matching (RTHM) uses a
similarity algorithm related to the Hausdorff
distance to enhance linkage peak resolution.
c
d
BACKGROUND Spike Trains
GL
REFERENCES
GL
1 second
Action potential occurrence times may be
extracted from multi-neuronal extracellular
recordings by a template matching process.
a, c RTHM correlogram and RTHM scatterplot for
two MCL units (cells 11-3 11-4). The density
peaks and bands at the arrows suggest
short-latency mutual excitatory interactions.
The void surrounding the zero latency line is an
artifact of extracting two unit trains from a
single-electrode recording. b, d Shuffled ISI
sequence versions of a,c with the central region
masked, to demonstrate the uniform RTHM
distributions for expressly uncorrelated units.
RTHM Correlogram
MCL
De la Luz M, Diaz E and Maldonado PE.
Simultaneous single unit recording in the
mitral cell layer of the rat olfactory bulb
under nasal and tracheal breathing. Biological
Research. 38 13-26, 2005. Freeman WJ. Mass
Action in the Nervous System. Academic
Press 1975. Rhoades BK. Excitatory Actions of
GABA in the Olfactory Bulb. Doctoral
Dissertation,University of California at
Berkeley, 1990. Rhoades BK, Kowalski JM, Mackey
HJ and Gross GW. Graphical measures of spike
train similarity and coupling based on
recurrence time history matching (RTHM).
Society for Neuroscience 24th Annual Meeting.
Miami Beach, FL, 1994. Shepherd GM. The
Synaptic Organization of the Brain. Oxford
University Press, 1979.
enhanced linkage peak
ML Cells With Strong Excitatory Linkages Evidence
Activity-Dependent Excitatory Gain
2
2
A RTHM scatterplot plots ranked local interspike
interval averages against RTHM cross-intervals.
Excitatory synaptic linkages appear as vertical
bands, with density reflecting activity-dependent
synaptic gain. Excitatory gain drops off at high
and low local firing rates.
1
1
METHODS Recording and Analysis
a
b
2
2
1
2
.
Analog Recording Selected 5-10 minute epochs
meeting the criteria for MCL activity were
recorded to analog cassette tapes (RCA
SCT-510). Spike Train Extraction One or more
single-unit spike trains were extracted from
analog recordings using a an adaptive
template-matching system (Cambridge Electronic
Design 1401/Spike-2). Correlational
Analysis Correlational analysis, including ISIHs
and RTHM plots, was performed on spike train
pairs (after Rhoades et al, 1994).
c
d
A unit spike train is a temporal series of
sequential occurrence times of action potentials
extracted for a single neuron.
excitatory linkage band
ACKNOWLEDGEMENTS
a, b RTHM correlogram and RTHM scatterplot for
two ML units (cells 9-1 9-2) with strong
excitatory linkages (arrows). c Extraction of a
linkage band (vertical lines) d ISIH for band in
c, showing reduction in linkage strength for
short ISIs/high firing rates (activity-dependent
gain).
This project was supported, in part, with funds
from NSF/CCLI grant DUE9950546 and an endowment
to Wesleyan College from the Munroe family of
Atlanta, GA.
time (5 msec ticks)