SPECTRAL AND CROSS-SPECTRAL ANALYSIS OF

CONDUCTANCE CATHETER SIGNALS

New Indexes for Quantification of Mechanical Dyssinchrony

Sergio Valsecchi

Medtronic Italia, Rome, Italy

Luigi Padeletti

University of Florence, Florence, Italy

Giovanni Battista Perego

Istituto Auxologico Italiano, Ospedale S Luca, Milan, Italy

Federica Censi, Pietro Bartolini

Dept Technologies and Health - Italian National Institute of Health, Rome, Italy

Jan J. Schreuder

Dept of Cardiac Surgery, San Raffaele Hospital, Milan, Italy

Keywords: Conductance catheter, spectral analysis, coherence function, heart failure, mechanical ventricular

dyssynchrony.

Abstract: We hereby present novel index to quantify ventricular mechanical dyssynchrony by using

spectral and cross-spectral analysis of conductance catheter volume signals. Conductance

catheter is a volume measurement technique based on conductance measurement: the

intraventricular volume, i.e. the time-varying volume of blood contained within the

heart cavity, is

estimated by measuring the electrical conductance of the blood employing a multi-pole catheter. Five

segmental volume signals (SV

, i=1,…5) can be acquired; total volume (TV) is estimated as the

instantaneous sum of the segmental volumes. We implemented classical time-domain dyssynchrony indexes

already utilized in conductance catheter signals analysis, and new frequency-domain indexes. Study

population consisted of 15 heart failure (HF) patients with left bundle branch block and 12 patients with

preserved left ventricular (LV) function. We found that spectral measures seem to out-perform classical

time-domain parameters in differentiating atrial HF patients from no-HF group. These findings encourage

the use of spectral analysis

to obtain crucial quantitative information from conductance catheter

signals.

1 INTRODUCTION

In a normal heart, mechanical activation of the

ventricles occurs in a coordinated manner and

depends on the rapid spread of electric signals via

specialized fibers (His-Purkinje system) which

branch out throughout the right ventricular (RV) and

left ventricle (LV) endocardium (Uhley, 1960).

When the activation is slowed-down or blocked,

ventricle activation and contraction become

dyssynchronous. Ventricular mechanical

dyssynchrony is most commonly identified clinically

by a prolonged QRS duration with left bundle-

branch block (LBBB) morphology on surface

437

Valsecchi S., Padeletti L., Battista Perego G., Censi F., Bartolini P. and J. Schreuder J. (2008).

SPECTRAL AND CROSS-SPECTRAL ANALYSIS OF CONDUCTANCE CATHETER SIGNALS - New Indexes for Quantiﬁcation of Mechanical

Dyssinchrony.

In Proceedings of the First International Conference on Bio-inspired Systems and Signal Processing, pages 437-444

DOI: 10.5220/0001060104370444

 SciTePress

electrocardiogram but can also be detected by

echocardiographic imaging of contraction timing.

Ventricular mechanical dyssynchrony plays a

regulating role already in normal physiology

(Brutsaert, 1987) but is especially important in

pathological conditions, such as hypertrophy (Villari

et al., 1996), ischemia (Heyndrickx and Paulus,

1990), infarction (Gepstein et al., 1998), or heart

failure (HF) (Nelson, 2000). Dyssynchrony

exacerbates heart failure (HF) in a variety of ways,

generating cardiac inefficiency as well as pathologic

changes at the biologic tissue, cellular, and

molecular levels. Currently, the conductance

catheter method has been extensively used to assess

global systolic and diastolic ventricular function.

More recently the ability of this instrument to pick-

up multiple segmental volume signals has been used

to quantify mechanical ventricular dyssynchrony.

Figure 1: The conductance catheter positioning inside the

left ventricle.

Conductance catheter was first introduced in

1981, by Baan and co-workers as a new volume

measurement technique based on conductance

measurement (Ban et al., 1981, Ban et al., 1984).

Intraventricular volume, i.e. the time-varying

volume of blood contained within the heart cavity, is

estimated by measuring the electrical conductance of

the blood employing a multi-pole catheter

(conductance catether, Figure 1). The conductance

catheter has 12 electrodes and should be positioned

along the long axis of the LV in such a way that the

electrode at the tip is situated within the apex and

the proximal one just above the aortic valve. A weak

alternating current (0.4 mA peak-to-peak, 20 kHz) is

induced between the two most distal and two most

proximal electrodes, in order to set up an electrical

field within the ventricular cavity. Other 6 electrodes

are used pair wise to measure segmental

conductance signals. Two electrodes are used to

record the intracardial ECG. A micromanometer

measures real-time LV pressure. The induced

voltage is then measured with six electrodes in

between, yielding 5 segmental voltages. Since the

conductance of the blood itself is constant

(neglecting long term changes in haematocrite) the

measured voltage will be proportional to blood

resistivity, and thus inversely proportional to the

conductance or amount of blood between the

measuring (voltage) electrodes. This method has

several advantages over other methods which

determine intra-ventricular volumes. The results are

obtained immediately, i.e. on-line, and precise

geometric assumptions regarding the ventricle or

labor-intensive analyses are not required. Recently,

Steendijk et al., first introduced time-domain

quantitative indexes of dyssynchrony based on

volume signals acquired with the conductance

catheter (Steendijk et al., 2004). Spectral analysis of

conductance catheter signals has not been tempted

yet. Frequency domain analysis has been extensively

used to characterize a number of physiological

signals, with promising results in terms of both

classification schemes (Schumann et al., 2002, 1

Zywietz et al., 2004, Severi et al., 1997) and

understanding of physiological mechanisms (Asyali

et al., 2007, Cerutti et al., 1988, Montano et al.,

2001). During ventricular dysfunction, segmental

ventricular volumes experience abnormal changes

which could result in unexpected spectral

components. Also, the asynchrony and

incoordination between ventricular segments, that

seem to be quintessential to ventricular dysfunction,

could be promisingly explored by cross-spectral

analysis. The coherence spectrum is a frequency

domain measure that may be used to make a

quantitative comparison between activity of two

heart regions. In the present study, coherence spectra

have been used to quantify the relation between

spectral components of ventricular volumes from

different regions. The coherence spectrum would

provide a measure of the synchrony and

coordination between ventricular sites, and thus be

indicative of the organization of electrical activity.

Such a measure would be a useful tool in the

BIOSIGNALS 2008 - International Conference on Bio-inspired Systems and Signal Processing

438

characterization and detection of synchronous

contraction. Coherence measurements may provide a

means to quantify the terms "synchronous" and

"dissynchronous" as applied to ventricular

contraction.

Aim of this paper is to characterize the

conductance catheter signals in the frequency

domain and to propose new indexes for ventricular

mechanical dyssynchrony quantification.

2 METHODS AND MATERIALS

2.1 Study Population

The study population consisted of 27 consecutive

patients with indications for electrophysiologic study

or device implantation: 15 HF patients with left

bundle branch block and 12 patients with preserved

LV function. Table 1 shows the clinical

characteristics of the study population. Age, sex and

QRS duration were similar between groups. Subjects

with a previously implanted device, valvular

insufficiency or stenosis were excluded from

analysis.

Table 1: Clinical characteristics of the study population.

non-HF group

(n=12)

HF group

(n=15)

Male gender, n 7 11

Age, years

67±14 68±6

Ischemic

Cardiomyopathy, n

- 7

NYHA Class -

3.1±0.5

Ejection Fraction, %

57±9 26±6*

QRS duration, ms

88±21 167±24*

p-values: * < 0.05

2.2 Experimental Protocols

A conductance catheter was placed in the LV via the

femoral artery, and a temporary pacing lead was

positioned in the right atrium. The conductance

catheter enables online measurement of 5 segmental

volume (SV

, i=1,…,5) slices perpendicular to the

LV long axis. We used 7-Fr combined pressure-

conductance catheters with 1-cm interelectrode

spacing (CD Leycom; Zoetermeer, The

Netherlands). The catheter was connected to a

Cardiac Function Lab (CD Leycom) for online

display and acquisition (sample frequency 250 Hz)

of segmental and LV total volumes (TV), LV

pressure, and ECG. TV was obtained as the

instantaneous sum of the segmental volumes.

Two stimulation protocols have been used, i.e.

during spontaneous ventricular activation and atrial

pacing. For this protocol, hemodynamic status was

evaluated using multiple parameters. Indices of LV

pressure, volume, and function were calculated and

averaged over 8 to 10 beats at end expiration from

the raw LV pressure and conductance volume data.

Sequences of 30 s, i.e. 40-50 consecutive non-

arrhythmic cardiac cycles at fixed heart rate induced

by atrial pacing (at 10 bpm above the sinus rate) and

steady-state conditions, were selected for off-line

analysis using custom-designed software.

2.3 Classical Dyssynchrony Parameters

Estimation

From the conductance catheter signals, we estimated

the following classical time-domain parameters:

mechanical segmental dyssynchrony (DYS), Internal

flow fraction (IFF), Mechanical Dispersion (DISP),

Cycle Efficiency (CE) and Time exceeding aortic

closure (TExAC). See appendix for more details.

2.4 Spectral and Cross-Spectral

Analysis

First we analysed the segmental and total volume

signals in the frequency domain. For the spectral

analysis, the periodogram of the signals was

estimated. To reduce spectral leakage a Hamming

window was applied after removal of the mean

value. The length of segments was 1000 samples and

a segment-overlap of 30% was used. Then we

divided the signal bandwidth in 4 frequency bands

(0-1 Hz, 1-5 Hz, 5-20 Hz and >20 Hz), and we

estimated the percentage powers (PP) and the peak

frequencies (PF) in each bands (PP

0-1Hz

, PP

1-5Hz

, PP

20Hz

, PP

>20Hz

and PF

0-1Hz

, PF

1-5Hz

, PF

5-20Hz

, PF

>20Hz

respectively).

The continually changing temporal or phase

relationship between two volume signals has been

quantified in the frequency domain by magnitude-

squared coherence (Ropella et al., 1989).

Magnitude-squared coherence (C(f)) between two

recordings is defined as

)()(

)(

fSfS

yyxx

Where x(t) and y(t) are two simultaneous recordings,

Sxy is the cross power spectrum between signals x

and y, and Sxx and Syy are the individual power

spectra for signals x and y, respectively. C(f) is a

measure of the linear relation between signals as a

SPECTRAL AND CROSS-SPECTRAL ANALYSIS OF CONDUCTANCE CATHETER SIGNALS - New Indexes for

Quantification of Mechanical Dyssinchrony

439

function of frequency, f, and is a real quantity with

value between zero and one. In other terms, C(f)

measures the constancy of the time delay (phase) at

a specific frequency between signals x and y. Two

linearly related signals (in the absence of noise) will

have a C(f) function equal to one at all frequencies

present in both signals, while two random,

uncorrelated signals will have a C(f) equal to zero at

all frequencies. Any linear operation (multiplication

by a constant or addition of a constant) on one or

both of the signals will not alter the C(f) between x

and y. However, additive, uncorrelated noise and

system nonlinearities will reduce C(f) for two

similar signals. C(f) may be estimated for sampled

data using a method of overlapped and averaged

FFT spectral estimates (Carter et al., 1973).

Basically, estimates of Sxx, Syy and Sxy are

determined using a periodogram technique (2048-

samples long window, overlap 512 samples), and

their estimates are then used in the definition of C(f).

The C(f) functions between each segmental volume

SVi and the TV have been estimated. A Total

Coherence function has been defined over the band

0-125 Hz by averaging the 5 C(f) functions. From

the Total Coherence function, 5 new frequency

domain indexes have been extracted:

- mean value of the Total Coherence over the band

0-125 Hz (Coh

Tot

)

- mean value of the Total Coherence from 0 to 1 Hz

(Coh

0-1Hz

- mean value of the Total Coherence from 1 to 5 Hz

(Coh

1-5Hz

- mean value of the Total Coherence from 5 to 20 Hz

(Coh

5-20Hz

- mean value of the Total Coherence from 20 to 125

Hz (Coh

>20Hz

2.5 Statistical Analysis

All data are presented as means±SD. Differences

between distributions were compared by a t-test for

Figure 2: Examples of SV and TV time series and power spectra, for a no-HF patinet and a HF one. Coherence function

between SV

and TV are showed, as well as the total coherence function.

No-HF patient

HF patient

Power spectrum

Coherence function Sv

-TV

Coherence function Sv

-TV

Total Coherence

180

200

150

170

[ml]

Total volume

Segment-1 volume

Total volume

Segment-1 volume

0 [s] 10

0 [Hz] 20

0 [Hz] 125 0 [Hz] 125

300

[ml

300

[ml

3x10

[ml

3x10

[ml

BIOSIGNALS 2008 - International Conference on Bio-inspired Systems and Signal Processing

440

Gaussian variables, and by Mann-Whitney

nonparametric test for nongaussian variables.

Statistical correlations between variables were tested

by least-squares linear regression. A P value < 0.01

was considered significant. We performed receiver-

operating characteristic (ROC) curve analysis to test

the diagnostic performance of the indexes to

discriminate the patient groups. Sensitivities and

specificities at the optimal cut-off point were

determined.

3 RESULTS

Example of the power spectrum of a TV signal are

showed in figure 2, for a HF patient and a no-HF

one. The coherence function between one SV and

the TV and the Total coherence are also showed.

The characteristics of the power spectrum of TV

signals are reported in Table 2 (similar results were

obtained for SVi signals, but were not reported). The

majority of the signal power is in the band from 1 to

5 Hz (programmed heart rate during acquisition

from 70 to 100 bpm). The components above 20Hz

are associated to less than 1% of the total signal

power. The frequency peak in the 0 – 1 Hz band

matches with the respiratory rate and the power in

this band seems higher in HF group. Table 3

summarizes the results of the comparison between

groups for all indexes considered in the analysis,

represented as mean ± standard deviation. Overall, 3

parameters permitted to discriminate the two groups

(p<0.01): Coh1-5, Coh5-20 and CE. Table 4 shows

the results of the ROC curve analysis. Sensitivity

and specificity for Coh1-5 are 0.67 and 0.92, those

obtained for Coh5-20 are 1.00 and 0.92 and those

relatve to CE are 0.80 and 0.83, respectively. In

Figure 3 the ROC curves are shown.

Table 2: Characteristics of the power spectrum of TV

signal.

no-HF HF p-values

0-1Hz

4.34±6.26 9.12±12.00 0.071

0-1Hz

0.41±0.17 0.45±0.12 0.488

1-5Hz

93.24±6.45 88.44±11.85 0.194

1-5Hz

1.52±0.20 1.44±0.13 0.301

5-20Hz

2.16±1.84 2.21±1.34 0.946

5-20Hz

8.37±0.88 7.63±0.80 0.036

>20Hz

0.25±0.25 0.23±0.15 0.814

>20Hz

37.59±5.12 44.14±7.59 0.013

4 DISCUSSION

Quantification of nonuniform mechanical function

and dyssynchrony may lead to a more complete

diagnosis of ventricular dysfunction (Schreuder wet

al., 1997, Schreuder et al., 2000). Moreover, it may

guide therapy, because patients with extensive

dyssynchrony are likely to benefit from

resynchronization therapy (Leclercq et al., 2002).

The visualization of mechanical dyssynchrony

provided by methods based on magnetic resonance

imaging and echocardiography, although further

emphasize the important role of mechanical

dyssynchrony in cardiac dysfunction, requires

laborious procedures and require substantial operator

interaction and expertise.

Table 3: Indexes of mechanical dyssynchrony in no-HF

and HF groups.

no-HF HF p-values

DYS, % 26.0±7.2 32.6±3.9 0.012

IFF, % 25.8±18.8 40.8±13.6 0.033

DISP, ms 23.4±16.4 35.6±13.2 0.068

CohTot 0.44±0.07 0.37±0.10 0.016

Coh0-1 0.63±0.19 0.51±0.18 0.099

Coh1-5 0.69±0.10 0.57±0.10 0.004*

Coh5-20 0.47±0.07 0.32±0.04 0.000*

Coh>20 0.43±0.08 0.37±0.12 0.041

CE 0.78±0.12 0.58±0.16 0.000*

TExAC, ms 6.9±8.8 15.7±10.5 0.016

*p<0.01

Figure 3: ROC curve analysis.

SPECTRAL AND CROSS-SPECTRAL ANALYSIS OF CONDUCTANCE CATHETER SIGNALS - New Indexes for

Quantification of Mechanical Dyssinchrony

441

Table 4: ROC curve analysis of the tested variables.

Area Under Curve

(95% CI)

p-value Cut-off

Coh1-5

0.81

(0.65-0.98)

0.006 0.57

Coh5-20

0.98

(0.94-1.02)

0.000 0.40

0.88

(0.75-1.00)

0.001 0.70

Recently, novel indexes were introduce to quantify

dyssynchrony based on volume signals acquired by

the conductance catheter during cardiac

catheterization (9). Such indexes were based on a

time-domain approach and provided additional, new,

and quantitative information on temporal and spatial

aspects of mechanical dyssynchrony.

To our knowledge, conductance catheter volume

signals have never been studied in the frequency

domain. Since dyssynchrony refers to the

organization of the mechanical contraction of the

ventricle, it is natural to investigate such a

phenomenon by spectral and cross-spectral analysis

of ventricular segmental movements. The frequency-

domain analysis can indeed discover particular

aspects of interaction between volume signals

beyond the temporal relationships.

Present analysis permitted to describe some

characteristics of the conductance-volume signals.

The frequency analysis evidenced the absence of

relevant components above 20 Hz: this result

corroborates the validation of segmental signals

acquisition obtained by comparison with cine-

computerized tomography (16), whose sampling rate

has approximately the same value. The amplitude of

the components in the range 0-1 Hz, attributable to

the respiratory artefact, resulted markedly higher in

HF patients, this may be due to the higher

(mechanical) cardio-pulmonary interaction or to an

altered vasovagal activity.

More interesting results have been obtained by

cross-spectral analysis. The spectral coherence

function provides a quantitative measure of that

temporal synchrony and coordination between

activities of ventricular regions. During synchronous

mechanical contraction, multiple sites are activated

in an coordinated manner, and the phase relation

between activity from two sites is relatively

unchanging, resulting in a high (close to 1)

coherence. When ventricular contraction is

dyssynchronous, the activity observed at one region

is likely to be unrelated to the activity observed at

other distant regions. Thus the coherence between

two such sites would be very low at all frequencies

due to a continually changing phase relation.

In the present study, the spectral coherence was

confirmed to be significantly greater for ventricular

contraction of no-HF patients than for HF ones. We

found that the most significant parameters in the

discrimination between HF and no-HF group were

Coh1-5 and Coh5-20, with the latter reaching a

sensitivity of 1 and a specificity of 0.92. Spectral

measures seem to out-perform classical time-domain

parameters (9) in differentiating atrial HF patients

from no-HF group.

Since no previous studies have been performed

on a similar topic, the frequency bands have been

chosen on empirical basis. The choice of the optimal

frequency bands in term of discriminating power

would require larger population and/or modelling of

ventricular contraction.

In conclusion, this paper encourages the use of

spectral analysis to obtain crucial quantitative

information from conductance catheter signals.

REFERENCES

Uhley H: Peripheral distribution of the canine A-V

conduction system; observations on gross

morphology. Am J Cardiol 5:688- 691, 1960.

Brutsaert DL. Nonuniformity: a physiologic modulator of

contraction and relaxation of the normal heart. J Am

Coll Cardiol 9: 341–348, 1987.

Villari B, Vassalli G, Betocchi S, Briguori C, Chiariello

M, and Hess OM. Normalization of left ventricular

nonuniformity late after valve replacement for aortic

stenosis. Am J Cardiol 78: 66–71, 1996.

Heyndrickx GR and Paulus WJ. Effect of asynchrony on

left ventricular relaxation. Circulation 81: 41–47, 1990.

Gepstein L, Goldin A, Lessick J, Hayam G, Shpun S,

Schwartz Y, Hakim G, Shofty R, Turgeman A,

Kirshenbaum D, and Ben-Haim S. A

Electromechanical characterization of chronic

myocardial infarction in the canine coronary occlusion

model. Circulation 98: 2055–2064, 1998.

Nelson GS, Curry CW, Wyman BT, Kramer A, Declerck

J, Talbot M, Douglas MR, Berger RD, McVeigh ER,

and Kass DA. Predictors of systolic augmentation

from left ventricular preexcitation in patients with

dilated cardiomyopathy and intraventricular

conduction delay. Circulation 101: 2703–2709, 2000.

Baan J, Jong TTA, Kerkhof PLM, et al. Continuous stroke

volume and cardiac output from intraventricular

dimensions obtained with impedance catheter.

Cardivascular Research 1981;15:328-334.

Baan J, van der Velde ET, de Bruin HG, et al. Continuous

measurement of left ventricular volume in animals and

humans by conductance catheter. Circulation

1984;70:812-823.

BIOSIGNALS 2008 - International Conference on Bio-inspired Systems and Signal Processing

442

Steendijk P, Tulner SAF, Schreuder JJ, et al.

Quantification of left ventricular mechanical

dyssynchrony by conductance catheter in heart failure

patients. Am J Physiol 2004;286:H723-H730

Schumann A, Wessel N, Schirdewan A, Osterziel KJ,

Voss A. Potential of feature selection methods in heart

rate variability analysis for the classification of

different cardiovascular diseases. Stat Med. 2002 Aug

15;21(15):2225-42.

Zywietz CW, Von Einem V, Widiger B, Joseph G. ECG

analysis for sleep apnea detection. Methods Inf Med.

2004; 43(1):56-9.

Severi S, Cavalcanti S, Avanzolini G. Heart rate

variability spectral indices for haemodynamic

classification of haemodialysis patients. Physiol Meas.

1997 Nov; 18(4):339-53.

Asyali MH, Berry RB, Khoo MC, Altinok A. Determining

a continuous marker for sleep depth. Comput Biol

Med. 2007 Apr 12;

Cerutti S, Alberti M, Baselli G, Rimoldi O, Malliani A,

Merri M, Pagani M. Automatic assessment of the

interaction between respiration and heart rate

variability signal. Med Prog Technol. 1988; 14(1):7-19.

Montano N, Porta A, Malliani A. Evidence for central

organization of cardiovascular rhythms. Ann N Y

Acad Sci. 2001 Jun;940:299-306. Review.

Ropella KM, Sahakian AV, Baerman JM, Swiryn S. The

coherence spectrum. A quantitative discriminator of

fibrillatory and nonfibrillatory cardiac rhythms.

Circulation 1989;80:112-9.

Carter C, Knapp CH, Nuttall B. Estimation of the

magnitudesquared coherence function via overlapped

fast Fourier Transform processing. IEEE Trans Audio

Electroacoustics. 1973;21:337-44.

Schreuder JJ, Steendijk P, Van der Veen FH, et al. Acute

and short-term effects of partial left ventriculectomy in

dilated cardiomyopathy: assessment by pressure-

volume loops. J Am Coll Cardiol 2000;36:2104-2114.

Schreuder JJ, van der Veen FH, van der Velde ET, et al.

Left ventricular pressure-volume relationships before

and after cardiomyoplasty in patients with heart

failure. Circulation. 1997;96:2978-2986.

Leclercq C, Kass DA. Retiming the failing heart:

principles and current clinical status of cardiac

resynchronization. J Am Coll Cardiol 39: 194–201, 2002.

APPENDIX

Mechanical dyssynchrony. At each time point, a

segmental signal is defined as dyssynchronous if its

change (i.e., dSV/dt) is opposite to the simultaneous

change in the total LV volume (dTV/dt). Segmental

dyssynchrony is quantified by calculating the

percentage of time within the cardiac cycle that a

segment is dyssynchronous. Overall LV

dyssynchrony (DYS) is calculated as the mean of the

segmental dyssynchronies. DYS may be calculated

within each specified time interval, i.e. during

systole and diastole, with systole defined as the

period between the moments of dP/dtmax and

dP/dtmin.

Internal flow. Nonuniform contraction and filling

is associated with ineffective shifting of blood

volume within the LV. This internal flow (IF) is

quantified by calculating the sum of the absolute

volume changes of all segments and subtracting the

absolute total volume change:

[

]

2//)(/)()(

∑

−= dttdTVdttdSVitIF

Note that dTV(t)/dt represents the effective flow into

or out of the LV. Thus IF measures the segment-to-

segment blood volume shifts, which do not result in

effective filling or ejection. Division by two takes

into account that any non-effective segmental

volume change is balanced by an equal but opposite

volume change in the remaining segments. IF

fraction (IFF) is calculated by integrating IF(t) over

the full cardiac cycle and dividing by the integrated

absolute effective flow.

Mechanical dispersion. In the HF patients, a

substantial dispersion is present in the onset of

contraction between the segments. This dispersion is

assessed by segmental lag times which are

determined by calculating the cross correlations

between TV(t) and SV(t) for all systolic time points

(i.e., between dP/dtmax and dP/dtmin). For each

segment the lag which produces the highest linear

correlation is determined. Mechanical dispersion

(DISP) is defined as 2 standard deviation of the

segmental lag times. Recently, new parameters have

been introduced to quantify LV dyssynchrony with

echocardiographic techniques. These indices can be

directly applied to conductance method.

Cycle Efficiency. Calculated as previously described

by the formula: CE=SW/[ LVP* LV volume], with

SW = stroke work, LVP = end-systolic – end-

diastolic LV pressure. This index quantifies

distortions in the shape of the pressure-volume

diagram. The calculation assumes that the optimal

contraction would have CE value near 1.0,

corresponding to a rectangular pressure volume

diagram. Decreases in cycle efficiency may be

caused by multiple factors including isovolumic

volume shifts as well as changes in afterload and

ventricular stiffness. Similarly, regional cycle

efficiency can be calculated from the most basal to

the most apical segmental volume signal plotted

against LV pressure. Differences in regional cycle

efficiency during isovolumic filling or emptying

may indicate inefficient patterns contraction or

relaxation due to dyssynchrony.

SPECTRAL AND CROSS-SPECTRAL ANALYSIS OF CONDUCTANCE CATHETER SIGNALS - New Indexes for

Quantification of Mechanical Dyssinchrony

443

Time exceeding aortic closure. In order to measure

diastolic dyssynchrony and specifically to quantify

LV contraction in diastolic phase, a new index was

proposed, quantitatively reflecting the whole

temporal amount spent by 12 LV segments in

contracting after aortic valve closure. Using strain

imaging that reflects myocardial deformation, the

time of strain tracing exceeding aortic valve closure

(ExcT) was measured in each segment as the

interval between the marker of aortic closure and the

nadir of the strain tracing. ExcT was considered 0

when the nadir of strain curve did not exceed aortic

valve closure. The overall time of strain exceeding

aortic valve closure (oExcT) was computed as the

sum of the 12 segmental ExcTs. The index may be

implemented in conductance method by considering

each segment presenting a systolic phase (negative

dSVi/dt) persisting during the phase of global

diastole (positive dTV/dt). oExcT is estimated as the

sum of these delays for all segments.

BIOSIGNALS 2008 - International Conference on Bio-inspired Systems and Signal Processing

444