ORIGINAL ARTICLES Interictal to Ictal Transition in Human Temporal Lobe Epilepsy Insights F

O RIGINAL A RTICLES

Interictal to Ictal Transition in Human Temporal Lobe Epilepsy:Insights From a Computational Model of

Intracerebral EEG

Fabrice Wendling,*Alfredo Hernandez,*Jean-Jacques Bellanger,*

Patrick Chauvel,?and Fabrice Bartolomei?

Summary:In human partial epilepsies and in experimental models

of chronic and/or acute epilepsy,the role of inhibition and the

relationship between the inhibition and excitation and epileptogen-

esis has long been questioned.Besides experimental methods carried

out either in vitro(human or animal tissue)or in vivo(animals),

pathophysiologic mechanisms can be approached by direct record-

ing of brain electrical activity in human epilepsy.Indeed,in some

clinical presurgical investigation methods like stereoelectroencepha-

lography,intracerebral electrodes are used in patients suffering from

drug resistant epilepsy to directly record paroxysmal activities with

excellent temporal resolution(in the order of1millisecond).The

study of neurophysiologic mechanisms underlying such depth-EEG

activities is crucial to progress in the understanding of the interictal

to ictal transition.In this study,the authors relate electrophysiologic

patterns typically observed during the transition from interictal to

ictal activity in human mesial temporal lobe epilepsy(MTLE)to

mechanisms(at a neuronal population level)involved in seizure

generation through a computational model of EEG activity.Intra-

cerebral EEG signals recorded from hippocampus in?ve patients

with MTLE during four periods(during interictal activity,just

before seizure onset,during seizure onset,and during ictal activity)

were used to identify the three main parameters of a model of

hippocampus EEG activity(related to excitation,slow dendritic

inhibition and fast somatic inhibition).The identi?cation procedure

used optimization algorithms to minimize a spectral distance be-

tween real and simulated signals.Results demonstrated that the

model generates very realistic signals for automatically identi?ed

parameters.They also showed that the transition from interictal to

ictal activity cannot be simply explained by an increase in excitation

and a decrease in inhibition but rather by time-varying ensemble

interactions between pyramidal cells and local interneurons project-

ing to either their dendritic or perisomatic region(with slow and fast

GABA

A kinetics).Particularly,during preonset activity,an increas-

ing dendritic GABAergic inhibition compensates a gradually in-

creasing excitation up to a brutal drop at seizure onset when faster

oscillations(beta and low gamma band,15to40Hz)are observed.

These faster oscillations are then explained by the model feedback

loop between pyramidal cells and interneurons targeting their peri-

somatic region.These?ndings obtained from model identi?cation in

human temporal lobe epilepsy are in agreement with some results

obtained experimentally,either on animal models of epilepsy or on

the human epileptic tissue.

Key Words:Human TLE—Hippocampus—Intracerebral EEG—

Neuronal population model—Parameter identi?cation—Ictogen-

esis mechanisms.

(J Clin Neurophysiol2005;22:343–356)

In human partial epilepsies and in experimental models

of chronic and/or acute epilepsy,the role of inhibition and the

relationship between the inhibition and excitation and epilep-

togenesis has long been questioned(Dalby and Mody,2001;

Dichter,1997).Recent advances have suggested that different

types of inhibition(depending on the location of synapses/

receptors,for example)are present in a given neuronal tissue

and that they may be differently involved in epileptic pro-

cesses.Particularly,in experimental focal models(kainate or

pilocarpine-treated rat),it has been shown that GABAergic

inhibition is impaired,but not uniformly:dendritic inhibition

is reduced whereas somatic inhibition is preserved(Cossart et

al.,2001;Houser and Esclapez,1996).This intact inhibition

could be due to selective survival of perisomatic inhibitory

interneurons,as shown in human epileptogenic tissue(scle-

rotic hippocampus in temporal lobe epilepsy[TLE])(Wittner

et al.,2001;Wittner et al.,2005).

Besides experimental methods carried out either in

vitro(human or animal tissue)or in vivo(animals),patho-

physiologic mechanisms can be approached by direct record-

ing of brain electrical activity in human epilepsy.Indeed,in

some clinical presurgical investigation methods like stereo-

electroencephalography(SEEG)(Bancaud and Talairach,

1973),intracerebral electrodes are used in patients with

drug-resistant epilepsy to directly record paroxysmal activi-

ties with excellent temporal resolution(in the order of1

millisecond).The study of neurophysiologic mechanisms

underlying such depth-EEG activities is crucial to progress in

the understanding of the interictal to ictal transition.Here,we

*Laboratoire Traitement du Signal et de L’Image,INSERM Universite′de

Rennes1,Campus de Beaulieu,Rennes Cedex,France;and?Laboratoire

de Neurophysiologie et Neuropsychologie,INSERM Universite′de la

Me′diterrane′e,Marseille Cedex,France

Address correspondence and reprint requests to F.Wendling,Laboratoire

Traitement du Signal et de L’Image Universite′de Rennes1,INSERM

U642Campus de Beaulieu,Bat.22-35042Rennes Cedex,France;

e-mail:fabrice.wendling@univ-rennes1.fr.

Copyright?2005by Lippincott Williams&Wilkins

ISSN:0736-0258/05/2205-0343

Journal of Clinical Neurophysiology?Volume22,Number5,October2005343

propose to perform this study using not only descriptive signal analysis techniques but also computational modeling,a research methodology that has been largely developed for the study of various(patho)physiologic mechanisms and that attracts growing interest in neuroscience.We address the question whether EEG dynamics observed during the transi-tion from interictal to ictal activity could result from variable global interactions between populations of pyramidal cells and inhibitory interneurons,and whether these interactions could be revealed by appropriate model-based processing of intracerebral EEG signals.Hence,we relate electrophysi-ologic patterns of seizure generation in human hippocampus to excitatory and inhibitory interactions in a neuronal popu-lation model of hippocampal EEG activity.

This type of computational model(also referred to as “neuronal rate model”or“neuronal population dynamics”; Wilson and Cowan,1972)has been demonstrated to produce realistic epileptiform activities(Wendling et al.,2000).Re-cently,it has provided a means for linking some pathophys-iologic characteristics of hippocampus at a neuronal popula-tion level to the fast oscillations observed at seizure onset at the EEG level(Wendling et al.,2002).

Here,we report results about identi?cation of the three main parameters of this model(namely excitation,slow dendritic inhibition,and fast somatic inhibition)from record-ings performed in patients with mesial TLE(inverse prob-lem).Parameter identi?cation(based on evolutionary algo-rithms)is performed to produce signals that match those actually recorded during interictal to ictal transition.From the analysis of parameters evolution during this transition,hy-potheses are generated about excitatory and inhibitory inter-actions that take place between subsets of cells represented within the modeled hippocampal neuronal population.In particular,we found that during preonset activity(a few tens of seconds before seizure),an increasing dendritic inhibition (produced by interneurons mediating slow GABA A inhibitory postsynaptic currents[IPSCs])compensates a gradually in-creasing excitation up to a brutal drop at seizure onset.Faster ictal oscillations(beta and low-gamma band)are then ex-plained by the model feedback loop involving pyramidal cells

and interneurons targeting their perisomatic region(mediat-ing fast GABA A IPSCs).As described in the discussion,these ?ndings obtained from model identi?cation in human TLE are in agreement with some experimental studies conducted in animal models of epilepsy.

MATERIALS AND METHODS Intracerebral EEG Signals Recorded From

Hippocampus in Patients With Mesial Temporal Lobe Epilepsy

Five patients undergoing presurgical evaluation of drug-resistant mesial temporal lobe epilepsy(MTLE)were selected.All patients had a comprehensive evaluation includ-ing detailed history and neurologic examination,neuropsy-chological testing,routine MRI,surface EEG,and SEEG.As illustrated in Fig.1,SEEG recordings were performed using intracerebral electrodes placed intracranially according to Talairach’s stereotactic method(Bancaud and Talairach, 1973).

These patients were selected for the present study because they satis?ed the following criteria:(1)seizures involved the mesial temporal regions at the onset;(2)MRI was normal or suggestive of hippocampal sclerosis(hip-pocampal atrophy and increased T2signal);(3)one electrode was placed in the hippocampus;and(4)the electrophysi-ologic pattern during the transition to seizure was character-ized by the emergence of a low-frequency,high-amplitude spiking activity followed by a so-called rapid discharge in the hippocampus.This pattern has been reported to be typical in MTLE(Bartolomei et al.,2004;Engel et al.,1989;Spencer et al.,1992;Velasco et al.,2000).

Analyzed signals were recorded on a128-channel Del-tamed system.They were sampled at256Hz and recorded

on FIGURE1.Depth electrodes implanted for SEEG in temporal lobe epilepsy.(a)coronal view of a preoperative plan in which electrode#1records electrophysiologic activity within the hippocampal head(internal contacts)and the middle part of the middle temporal gyrus(external contacts)and in which electrode#2records the activity within the entorhinal cortex (internal contacts)and the anterior part of the inferior tempo-ral gyrus(external contacts).Generally,five to eight electrodes are used to spatially sample the temporal region.The implan-tation accuracy is per-operatively controlled by telemetric x-ray imaging.A postoperative computed tomography(CT) scan without contrast is then used to verify the absence of bleeding and the precise location of each recording lead. Intracerebral electrodes were then removed and an MRI per-formed,permitting visualization of the trajectory of each elec-trode.Finally,CT-scan/MRI data fusion can be performed to anatomically locate each contact along the electrode trajec-tory.(b)Electrodes include10to15contacts2mm long,0.8 mm in diameter,and1.5mm apart.

F.Wendling et al.Journal of Clinical Neurophysiology?Volume22,Number5,October2005

a hard disk(16bits/sample)using no digital?lter.Two hardware?lters are present in the acquisition procedure.The ?rst one is a high-pass?lter(cut-off frequency equal to0.16 Hz at–3dB)used to remove very slow variations that sometimes contaminate the baseline.The second one is a ?rst-order low-pass?lter(cut-off frequency equal to97Hz at –3dB)to avoid aliasing.

Definition of Periods of Interest

For the?ve patients,signals recorded from the hip-pocampus during the transition from interictal to ictal activity were visually analyzed.In each patient,four seg-ments of EEG activity(10seconds duration)were de?ned at different stages of this transition using visual criteria (Figs.2and3),as follows.During the interictal period (“intICTAL”segment,chosen1minute before seizure onset at least),EEG signals do not exhibit high amplitude transient spikes.The preonset period(preceding seizure onset)is marked by the appearance of high amplitude spikes(“preONSET”segments,from10to50seconds before seizure onset).The electrical onset of the seizure (“ONSET”segment)is characterized by the appearance of the rapid discharge re?ecting faster oscillations,typically in the beta and low-gamma frequency bands(15to40Hz). Finally,as seizure develops,the rapid discharge slows down and changes into a more rhythmic activity(“ICTAL”segment),typically in the theta or alpha frequency band(4 to10Hz).Computational Model of Hippocampus Neuronal Population and Simulation of EEG Signals

A macroscopic approach(neuronal population,adapted to the macroscopic size of SEEG electrodes)was used to build the EEG model,which has been detailed in previous reports(Wendling et al.,2002).Brie?y,this model is based on the global cellular organization of the hippocampus. Firstly,recurrent excitatory connections from pyramidal cells to pyramidal cells have been demonstrated to occur in CA1 (Thomson and Radpour,1991;Whittington et al.,1997). Secondly,a series of studies based on a variety of techniques (Miles et al.,1996)demonstrated that there are two types of GABA A synaptic responses in CA1pyramidal neurons:a fast one near the soma and a slow one at the dendrites.The?rst one—GABA A,fast—is a rapidly activated and decaying IPSC mediated by somatic synapses and the second one—GABA A,slow—is a slowly rising and decaying IPSC mediated by dendritic synapses.More recent works(Banks et al.,1998; White et al.,2000)suggested that two separate classes of interneurons(possibly basket cells and interneurons in stra-tum lacunosum-moleculare respectively called,for simplic-ity,GABA A,fast interneurons and GABA A,slow interneurons) give rise to these two IPSCs.Moreover,as suggested in Banks et al.(2000),both classes interact:GABA A,slow cells inhibit not only pyramidal cells but also GABA A,fast inter-neurons.

The model was designed to represent this functional organization of interacting subsets of principal cells and

inter-FIGURE2.Intracerebral EEG signals recorded from hippocampus during the transition from interictal activity to ictal activity in five patients with mesial temporal lobe epilepsy.Solid line rectangles correspond to segments of EEG signal(10seconds duration)on which model parameters were identified.intICTAL,interictal activity;preONSET,mixing of interictal activity and high-amplitude spikes that arise before seizure onset;ONSET,rapid discharge occurring at the electrical onset of the seizure and in which faster oscillations are observed;ICTAL,activity observed after the rapid discharge period as seizure develops.intICTAL segments were chosen1minute(at least)before seizure onset(note that high-amplitude epileptic spikes are not present in these segments). Journal of Clinical Neurophysiology?Volume22,Number5,October2005Interictal to Ictal Transition in TLE

neurons,as summarized in Fig.4a.It consists in three subsets of neurons,namely the main cells (i.e.,pyramidal cells),the slow dendritic-projecting inhibitory interneurons (GABA A,slow recep-tors),and the fast somatic-projecting inhibitory interneurons (GABA A,fast receptors).Interneurons receive an excitatory input (AMPA receptor-mediated)from pyramidal cells.The in?uence from neighboring or more distant populations is represented by an excitatory input n ?t ?(modeled by a positive mean gaussian white noise)that globally describes the average density of afferent action potentials.The model output corresponds to the postsynaptic activity of the subset of pyramidal cells that mainly contributes to the EEG signal (?eld potential).

In each subset,a linear transfer function is used to trans-form the average presynaptic pulse density of afferent action potentials (the input)into an average postsynaptic membrane potential (the output).This transfer function is of order 2.The

associated impulse response h EXC ?t ??EXC.a .t .e ?at ,h SDI ?t ??SDI.b .t .e ?bt ,and h FSI ?t ??FSI.g .t .e ?gt ,respectively,deter-mines the excitatory,slow dendritic inhibitory,and fast somatic inhibitory average postsynaptic membrane potentials,where EXC,SDI,and FSI represent the synaptic gains (Fig.4c).In this work,these key parameters are automatically identi?ed from real EEG signals recorded in the hippocampus.

The inhibitory feedback loop from the subset of fast somatic-projecting inhibitory interneurons uses a faster impulse response h FSI ?t ?(i.e.,producing faster IPSP)than h SDI ?t ?.Here,the important parameter is the average somatic time constant 1/g,which was chosen lower than the average dendritic time constant 1/b,consistently with data provided in Traub et al.(1999a)(see model parameter values given in Table 1as well as average postsynaptic potentials displayed in Fig.

4c).

FIGURE 3.(a )Intracerebral EEG sig-nals recorded from hippocampus during the transition from interictal activity to ictal activity in patient 2and (b )time-frequency distribution of its energy (spectrogram,red color indicates higher energy).This se-quence of electrophysiologic pat-terns is characteristic of the hip-pocampus activity observed in temporal lobe epilepsy of mesial or-igin.(c )Normalized power spectral densities computed on segments of EEG activity during the interictal pe-riod (intICTAL),before the begin-ning of seizure (preONSET),during the rapid discharge at seizure onset (ONSET),and finally after the rapid discharge (ICTAL)during seizure spread show that frequency con-tents of intICTAL and preONSET segments are quite similar.The ON-SET period is marked by the redistri-bution of signal energy in higher frequency band (15to 30Hz,in this case)whereas the ICTAL period is characterized by the appearance of a narrow band activity (around 5Hz,EEG theta band).

F.Wendling et al.Journal of Clinical Neurophysiology ?Volume 22,Number 5,October 2005

In each subset,in turn,a static nonlinear function (asym-metric sigmoid curve S ?v ??2e 0??1?e r ?v 0?v ??)is also used to model threshold and saturation effects in the relationship be-tween the average postsynaptic potential of a given subset and the average pulse density of potentials ?red by the neurons.

Finally,interactions between main cells and local neu-rons are summarized in the model by seven connectivity constants C 1to C 7,which account for the average number of synaptic contacts.

Each impulse response h EXC ?t ?,h SDI ?t ?and h FSI ?t ?intro-duces a pair of ?rst-order ordinary differential equation.Consequently,the model can be represented by a set of 14?rst-order ordinary differential equations,which can be re-duced to an equivalent set of 10equations (Wendling et al.2002).This set must be solved by numerical integration methods adapted to its stochastic nature.We used the Euler method with a time step equal to 1/256milliseconds (corre-sponding to the sampling period of real signals).

Model output can be represented as:

X s ?Mn r ,P

where M denotes the function operated by the model to produce Xs (the simulated EEG signal)from a realization n r of input noise n(t)and for given vector P of model

parameters.

FIGURE 4.Neuronal population model based on the cellular organization of the hippocampus.(a )Schematic representation.A whole population of neurons is considered inside which a subset of principal cells (pyramidal cells)project to and receive feedback from other local cells.Input to interneurons is excitatory (AMPA receptor-mediated).Feedback to pyramidal cells is either excitatory (recurrent excitation)or inhibitory (dendritic-projecting interneurons with slow synaptic kinetics—GABA A,slow —and somatic-projecting interneurons (gray rectangle)with faster synaptic kinetics—GABA A,fast ).As described in Banks et al.(2000),dendritic interneurons project to somatic ones.(b )Corresponding block diagram representation.In each subset,the average pulse density of afferent action potentials is changed into an average inhibitory or excitatory postsynaptic membrane potential using a linear dynamic transfer function of impulse response h EXC ?t ?,h SDI ?t ?,and h FSI ?t ?while this potential is converted into an average pulse density of potentials fired by the neurons using a static nonlinear function (asymmetric sigmoid curve,S(v )).The subset of somatic-projecting interneurons (gray rectangle)receives input from both subsets of pyramidal and dendritic interneurons.One of the model outputs represents the summated average postsynaptic potentials on pyramidal cells.It reflects an EEG signal.The three main parameters of the model respectively correspond to the average excitatory synaptic gain (EXC),to the average slow inhibitory synaptic gain (SDI),and to the average fast inhibitory synaptic gain (FSI).These three parameters are automatically identified.Other parameters are detailed in Table 1.(c )Time-course of average postsynaptic membrane potentials:excitatory,slow inhibitory,and fast inhibitory,respectively,obtained from h EXC ?t ?,h SDI ?t ?,and h FSI ?t ?for standard values of EXC (3.25mV),SDI (22mV),and FSI (10mV).

Journal of Clinical Neurophysiology ?Volume 22,Number 5,October 2005Interictal to Ictal Transition in TLE

Identification of Model Parameters EXC,SDI, and FSI Using Evolutionary Algorithms Problem Statement

Model-based interpretation requires identi?cation of model parameters that can be formalized as an optimization problem(Hernandez et al.,2002).Because model parameters are based on physiologic considerations,hypotheses for the interpretation of real EEG activity could be obtained from the optimal set of parameters P???EXC,SDI,FSI?,i.e.,the one that minimizes an error function??F?o,F?S?P??,where F?o and F?S?P?are estimators of a feature vector F,respectively, computed on the observed EEG(X O)and synthesized EEG (X S)using parameter vector P.

In the present study,F was primary based on spectral properties of the EEG signal.More precisely,the power within speci?c frequency bands B1(0–4Hz,delta),B2 (4–12Hz,theta and alpha),and B3(12–64Hz,beta and gamma)was retained as three?rst features.Because these spectral features did not distinguish well between interictal activity and interictal activity mixed with high-amplitude spikes,we added a fourth feature based on a morphologic criterion indicating the presence of sporadic high extrema values.It is simply de?ned as the difference between the?rst and last?-order quantiles.The four features were computed on simulated and real signals normalized by their standard deviation(this provides invariance with respect to different amplitude scale factors in the model and in reality).

In the expression of the error function,we use an estimation F?S?P?instead of the exact value F S?P?that cannot

be analytically computed as a function of P because of the nonlinear nature of the model.This estimation is computed over a?nite duration?and consequently is dependant on realization n r of model input noise.

We are facing an optimization problem presenting multiple local optima in which the error function(1)is not deterministic and(2)is not computable(and thus not differ-entiable).Consequently,deterministic optimization algo-rithms,such as those based on gradient descent,do not apply.

Two types of optimization methods are well suited to this problem:combinatory and exhaustive search methods on the one hand and stochastic search methods on the other hand.As the former present prohibitive computational costs, we used a particularly interesting family of stochastic search

TABLE1.Model Paremeters,Interpretation and Values Used to Produce EEG Signals With the Model

Parameter Interpretation Value

EXC Average excitatory synaptic gain Determined from identi?cation procedure

SDI Average inhibitory synaptic gain(slow dendritic inhibition

loop)

Determined from identi?cation procedure

FSI Average inhibitory synaptic gain(fast somatic inhibition

loop)

Determined from identi?cation procedure

1/a Average time constant of excitatory postsynaptic

potentials at

the dendrites of pyramidal cells

a?100s–1

1/b Average time constant of inhibitory postsynaptic

potentials at

the dendrites of pyramidal cells

b?50s–1

1/g Average time constant of inhibitory postsynaptic

potentials at

the soma of pyramidal cells

g?350s–1

C1,C2Average number of synaptic contacts in the feedback

excitatory loop

C1?C,C2?0.8C(with C?135)

C3,C4Average number of synaptic contacts in the slow

feedback inhibitory loop

C3?C4?0.25C

C5,C6Average number of synaptic contacts in the fast feedback

inhibitory loop

C5?C6?0.1C

C7Average number of synaptic contacts in the connection

between slow

and fast inhibitory interneurons

C7?0.8C

v0,e0,r Parameters of the asymmetric sigmoid function S

(transforming an average

PSP into an average density of action potentials)

v0?6mV e0?2.5s–1r?0.56mV–1

n(t)Excitatory input noise(positive mean gaussian white

noise)

Mean?90pulses?s–1SD?30pulses?s–1

Standard values(except g,C5,C6,and C7)were established in(Jansen and Rit,1995).Main parameters(EXC,SDI,and FSI)are automatically identi?ed from real intracerebral EEG signals.

F.Wendling et al.Journal of Clinical Neurophysiology?Volume22,Number5,October2005

methods,known as evolutionary algorithms(EAs).EAs have been the subject of intensive research during the last decade and have shown to be useful in the solution of hard identi?-cation problems,including different biomedical applications (Hernandez et al.,2002;Pena–Reyes and Sipper,2000).The optimization procedure using EAs is detailed in appendix. Variance of Identified Parameters and Significance of Parameter Changes

To evaluate the dispersion of triplets of identi?ed val-ues,the whole identi?cation procedure was applied30times to the same segment,leading to a set of30solutions for the optimal parameter set P???EXC,SDI,FSI?.The distribu-tions of parameter EXC,SDI,and FSI values obtained for the thirty iterations were represented in a so-called boxplot form for statistical analysis and comparison.This compact graph-ical display provides a way to picture how parameter values are distributed relative to de?ned percentiles,to identify if a distribution is skewed and if outliers or unusual data values are present.They are used to graphically determine if signif-icant statistical differences do exist between sets of parameter values.This graphical test consists in comparing boxes rep-resenting the lower and upper percentiles.If boxes do not overlap,parameter values are signi?cantly different.

RESULTS

Frequency Content of Segments of EEG Activity Chosen During intICTAL,preONSET, ONSET,and ICTAL Periods

For each patient,segments of EEG activity chosen during the interictal period(“intICTAL”segment),during the period preceding the onset rapid discharge(“preONSET”segment),during the rapid discharge itself that typically occurs at seizure onset(“ONSET”segment),and during the period that follows this rapid discharge as seizure develops (“ICTAL”segment)were analyzed for their frequency con-tent.These segments are shown in Fig.2.Normalized power spectral densities(PSDs)were computed using the standard periodogram method over10-second-duration segments in both the real and simulated case.PSDs were found to be very typical with respect to the analyzed period.The frequency content of intICTAL and preONSET segments was found to be relatively similar.Indeed,high-amplitude sporadic spikes that appear during the preONSET period do not affect sig-ni?cantly the global spectral energy distribution.For the?ve patients,the?rst major change in PSDs is observed at seizure onset,during the rapid discharge,which is marked by the redistribution of signal energy in a higher frequency band, typically from15to40Hz(EEG beta and low-gamma bands).The second major change corresponds to the slow-down of the rapid discharge during the ictal activity,charac-terized by a transition from broadband to narrowband activity ranging from3to10Hz(mostly within the EEG theta band). Results are exempli?ed in Fig.3,which also provides a time-frequency representation(spectrogram,Fig.3b)of the SEEG signal recorded during the transition from interictal to ictal activity in patient2(Fig.3a).In this case,the rapid discharge observed at seizure onset is relatively stationary with a maximum energy located in a frequency band ranging from20to30Hz.One can also notice the periodical aspect (around5Hz)of the signal recorded during ictal activity(Fig. 3c,bottom right).

Identification of Model Parameters and Simulation of EEG Activity

For each patient,the model parameter identi?cation procedure was?rstly applied on each segment of EEG activ-ity,to?nd appropriate parameter setting for the evolutionary algorithm.Several tests were performed to analyze the mean error value for each generation and the convergence rate of the EA,as a function of the main algorithm parameters.The best results were obtained with a mutation probability of p m ?0.2and a crossover probability of p c?0.9.The error function was computed for each individual by comparing spectral and morphologic features computed on simulated and real signal segments,as described in Materials and Methods.The size of the population has been?xed to200 individuals.The stopping criterion of the EA is met if no solution improvement is found during10consecutive gener-ations,or if the maximum number of iterations(?xed to200) is reached.The best individual of the last population was chosen as the identi?ed solution,leading,in each case,to a triplet of values for the three model parameters:EXC(exci-tation),SDI(slow dendritic inhibition),and FSI(fast somatic inhibition).

Some identi?cation results found during this step are presented in Fig.5,which displays four examples of simu-lated signals identi?ed from real EEG signals along with their corresponding PSDs and error values.It can be observed that identi?ed parameters lead the model to generate EEG signals that are close(in terms of morphologic and spectral content) to real EEG signals.One can notice that signals produced by the model are very realistic:waveforms are qualitatively similar to real ones and PSDs are quantitatively comparable (as exempli?ed by epileptic spikes in Fig.5,preONSET period).At this step,we also observed some common ten-dencies(among patients)in the evolution of model parame-ters for transitions of dynamics.To check whether these tendencies were signi?cant or only due to the variability of identi?ed parameters related to the randomness inherent to this identi?cation procedure and to the randomness of the observed data,we conducted an analysis of the dispersion of estimated parameters EXC,SDI,and FSI for each type of activity in each patient.

Dispersion of Identified Parameters

The identi?cation procedure was repeated thirty times with random initialization on each segment of activ-ity leading,in each case,to thirty triplets of values(EXC, SDI,FSI)for each patient(P1to P5)and for each period (intICTAL,preONSET,ONSET,ICTAL).These results are displayed in Fig.6,in the compact form of so-called boxplots(see Materials and Methods,“Variance of Iden-ti?ed Parameters”).Taken as a whole,the evolution of

Journal of Clinical Neurophysiology?Volume22,Number5,October2005Interictal to Ictal Transition in TLE

parameters (with respect to considered period of time)was found to be partly reproducible from patient to patient.One can also notice that the height of the boxes that represent the dispersion of the identi?ed EXC,SDI,and FSI parameters was small in most cases (except for parameter FSI during the preONSET period;see explanation below).This means that the identi?cation procedure generally converged to similar triplets of values.Indeed,detailed examination showed the error (quanti?ed using the error function)between the segments of activity simulated for these triplets and the real ones was low.The variance of the identi?cation error was also found to be low over each analyzed period.This is shown by the values of mean error and its standard deviation computed for all segments of a given activity (intICTAL,??0.00092?0.0052;preONSET:??0.013?0.0095;ON-SET:??0.013?0.0082;ICTAL:??0.027?0.016).Note also that in the examples previously provided in Fig.5,the error between synthesized and real activities corresponds to the above average errors.In each patient,the distribution of parameter FSI was found to be slightly broader when iden-ti?ed on the preONSET segment.During this period where parameter SDI increases,higher variability of the FSI param-eter estimation might be explained by the fact that fast inhibitory processes are depressed by slow ones in the model and that model output becomes less sensitive to FSI param-eter variations (see “Computational Model of Hippocampus Neuronal Population”in Material and Methods).Finally,one can also notice that only few outliers (with respect to the 30trials)are present in these boxplots.This means that the identi?cation procedure was able to converge to a reproduc-ible and small region in the parameter space for each period of interest and for each

patient.

FIGURE 5.Examples of simulated segments of EEG activity after automatic identification of model parameters from displayed segments of real EEG activity.Corresponding power spectral densities (solid line:real signal,dotted line:simulated signal).EXC,excitation;SDI,slow dendritic inhibition;FSI,fast somatic inhibition (see Table 1for details);?,error between simulated and real segments (see Materials and Methods for detailed description of error function).

F.Wendling et al.Journal of Clinical Neurophysiology ?Volume 22,Number 5,October 2005

Evolution of Model Parameters With Respect to Transitions

The evolution of model parameters with respect to transitions is given in Table 2.For the ?ve patients,an increase of parameter EXC was observed during the transi-tion from the intICTAL to the preONSET period and from the preONSET to the ONSET period.This parameter stayed either constant (P1)or increased (P2to P5)during the following transition (ONSET to ICTAL).Regarding param-eter SDI,evolution was found to be similar in the

?ve

FIGURE 6.Parameter identification results for the 20presegmented EEG signals (five patients,P1to P5,during four different EEG activities).Boxplots are calculated from 30independent realizations of the identification process for each segment of real activity.Results show the stability of the identified solutions and depict a global tendency in the evolution of parameter values according to the EEG activity.

Journal of Clinical Neurophysiology ?Volume 22,Number 5,October 2005Interictal to Ictal Transition in TLE

patients.An increase was?rst always observed from intICTAL to preONSET.Then,the transition from the preONSET to the ONSET period was marked by a noticeable decrease of parameter SDI.Finally,this parameter was found to signi?cantly reincrease during the transition from ONSET to ICTAL but without reaching again its initial value ob-served during the intICTAL period.Finally,regarding param-eter FSI,no signi?cant change was measured from intICTAL to preONSET and from preONSET to ONSET.During the transition from ONSET to ICTAL period,we observed a noticeable decrease of parameter FSI.

Invariance of Parameter Evolution With Respect to Transitions

To determine whether reproducible changes do exist in the evolution of EXC,SDI,FSI parameters with respect to the type of transition,we compared results obtained in the?ve patients.These are reported in Fig.7a,which gives the evolution of each parameter,for each transition and for each patient(?,signi?cant increase;–,signi?cant decrease;?,no signi?cant change).We also averaged the identi?ed parame-ter values of a given run over the?ve patients to obtain30 mean values for each parameter and for each period.Figure 7b displays the distribution of these mean values in the form of boxplots.Results show that signi?cant changes(no overlap between boxes)do exist and con?rm the scenario described hereafter.Regarding parameter EXC(average excitatory syn-aptic gain),a signi?cant increase was observed between the intICTAL and preONSET period and between the preONSET and the ONSET period.Hence,when epileptic sporadic spikes as well as fast oscillations at seizure onset are ob-served,an increase of the average excitatory postsynaptic potential amplitude in the projections from pyramidal cells to interneurons as well as in the loop representing recurrent excitation is concomitantly observed in the model.

As far as model parameter SDI is concerned,three major changes were observed.First,parameter SDI increased from intICTAL to preONSET period.We noticed that this signi?cant increase might be due to the aforementioned increase of excitation and could be related to a compensatory

mechanism.Second,parameter SDI also signi?cantly de-creased when identi?ed over the ONSET period.Indeed, during fast oscillations activity at seizure onset,parameter SDI reached its lowest value in the?ve studied cases.Third, a reincrease of parameter SDI is observed during ICTAL activity.As discussed below,this result suggests that

tran-FIGURE7.(a)Comparison of results obtained in the five patients about the evolution of EXC,SDI,FSI parameters with respect to the type of transition(?,significant increase;–, significant decrease;?,no significant change).Similar changes were found in the five studied cases(gray boxes).(b) Identified parameter values of a given run of the identification procedure were averaged over the five patients.Distributions of averaged values in the form of boxplots show that signifi-cant changes(no overlap between boxes)do exist in the evolution of model parameters with respect to the type of activity(interictal,preonset,onset,and ictal).

TABLE2.Average Values and Standard Deviation of Identified Model Parameters

EXC SDI FSI InterICTAL PreONSET ONSET ICTAL InterICTAL PreONSET ONSET ICTAL InterICTAL PreONSET ONSET ICTAL

P1 3.8?0.4 5.3?0.78.6?0.98.3?0.913.9?2.228.4?7.0 3.3?0.89.6?0.67.7?4.415.9?4.015.3?2.13.1?0.3 P2 3.8?0.4 4.7?0.68.2?1.19.96?0.0813.3?2.021.9?4.5 2.6?0.917.0?0.7 6.3?4.212.8?5.415.6?2.29.0?1.1 P3 4.4?0.4 4.9?0.58.8?0.710.0?0.418.0?2.028.1?3.7 2.6?0.714.7?2.312.1?4.1 4.8?6.117.0?2.17.1?5.7 P4 4.6?0.2 5.1?0.67.6?1.19.4?0.717.2?1.226.4?4.8 2.4?2.59.4?0.712.8?2.514.1?3.816.9?2.12.5?0.7 P5 3.9?0.2 5.2?0.4 6.3?1.59.8?0.815.1?1.625.4?2.4 1.8?4.311.0?3.09.1?3.214.0?4.314.3?2.51.6?4.1

Average values and standard deviation of identi?ed model parameters EXC,SDI,FSI for each one of the four considered periods of time(interictal, preonset,onset,and ictal)and for each patient(P1to P5).

F.Wendling et al.Journal of Clinical Neurophysiology?Volume22,Number5,October2005

sient failure of GABA A,slow interneurons can occur during the transition to seizure.

Finally,concerning parameter FSI,two major observa-tions were made.Over the ONSET period,in contrast to slow inhibition,FSI values were found either to be stable or to increase with respect to the preONSET period.During the transition from ONSET to ICTAL period,parameter FSI signi?cantly decreased.In the model framework,these results suggest that fast oscillations are related to the fast feedback loop activity of GABAergic interneurons targeting pyramidal cells in their perisomatic region(GABA A,fast kinetics)and that the slowing down of the onset rapid discharge is related to an exhaustion of this fast inhibition.

DISCUSSION

Different modeling strategies have been proposed to relate EEG signals with oscillations in neuronal populations. Besides modeling techniques lying at the cellular and net-work levels(Traub et al.,1999b),another approach lying at a higher level of organization,i.e.,the population level,starts from the fact that neurons form populations.The resulting EEG is a re?ection of ensemble dynamics rising from inter-connections between a small number of subpopulations of pyramidal cells on the one hand,and interneurons on the other hand.This approach was initially proposed by Freeman and coworkers(1978),who made substantial progress in the understanding of perceptual processing in the olfactory sys-tem.Similar ideas developed in the same time by Lopes da Silva et al.(1974)led to the development of a lumped-parameter population model used to explain the alpha rhythm of the EEG and more recently to study the mechanisms underlying transitions between spike and wave discharges in nonconvulsive epilepsy(Suffczynski et al.,2004).

On the basis of experimental results about cellular organization in the hippocampus,we followed a similar approach to build a physiologic-relevant neuronal population model.Main features include an excitatory loop and two parallel inhibitory circuits that represent the two separate populations of GABAergic interneurons that contact anatom-ically segregated postsynaptic receptors(dendritic/somatic region,GABA A,slow/GABA A,fast).As in any model,certain issues are simpli?ed.The major simpli?cation of our model is related to its macroscopic level.The reason for this choice is twofold.First,the neuronal population level is adapted to the nature of our observations:simulated signals can be directly compared with real intracerebral EEG signals recorded with macroelectrodes in human temporal lobe epilepsies.Second, we wanted to determine to what extent ensemble interactions between main cells and interneurons within hippocampal neuronal populations can explain dynamics and transitions of dynamics observed in real EEG signals.

Results showed that the model can produce various types of signal dynamics that were qualitatively and quanti-tatively similar to those observed in real depth-EEG signals recorded in human hippocampus over interictal and ictal periods.In?ve patients,the identi?cation procedure led to reproducible modi?cations about model parameter values of excitation,slow dendritic inhibition,and fast somatic inhibi-tion with respect of the analyzed type of EEG activity (interictal,preonset,onset rapid discharge,and ictal).In the following,these results are discussed with respect to current hypotheses about possible mechanisms involved in epilepto-genesis or ictogenesis.

Increased Excitation From Interictal to Preonset and Ictal Activity

Causes of hyperexcitability of the epileptic tissue have been largely discussed during the past decades(Jefferys, 2003).In the hippocampus,enhanced excitability may be caused by intrinsic and/or extrinsic factors.In the present study based one a single population,parameters changes may only be related to intrinsic mechanisms although we are aware that exogenous factors such as the enhancement of the direct entorhinal cortex input to CA1can play a role in excitability changes in the hippocampus,as reported in(Wu and Leung,2003).Results obtained from automatic identi?-cation of model parameters suggest an increase in the average excitatory synaptic gain in feedback loops that link pyramidal cells to local cells within the considered neuronal population during the transition from incterictal activity to preonset activity(interictal and sporadic spikes)and from preonset activity to ictal activity(beta and gamma oscillations at seizure onset and theta-like ictal activity).Although several studies performed in experimental models of focal epilepsy (Bernard et al.,2001;Gorter et al.,2002)or performed in slices from human brain(Schwartzkroin,1994)have already demonstrated that epileptic discharge patterns are correlated with prolonged and/or enhanced excitatory postsynaptic po-tentials,very few studies have attempted to relate them to gradual changes observed at the EEG level during the tran-sition to seizure,as done here through the model of the human hippocampus.Indeed,we infer from the model that a sustained excitation enhancement constitutes a necessary(but not suf?cient)condition to explain the appearance of interic-tal sporadic spikes and that the excitation level must stay high to get fast oscillations and?-like ictal activity on the EEG. Increase and Decrease of Slow Dendritic Inhibition From Interictal to Seizure Activity Another striking result of this study is related to the evolution of slow inhibition mediated by GABAergic inter-neurons targeting the dendrites of pyramidal cells (GABA A,slow kinetics,as represented in the model).Indeed, model-based interpretation of slow dendritic inhibition evo-lution suggests that it signi?cantly increases over the preonset period compared with interictal and abruptly decreases at seizure onset when faster oscillations are re?ected on the EEG.At?rst sight,the reinforcement of inhibition may appear as paradoxical because increased inhibition could be supposed to prevent the occurrence of seizures.We inter-preted this phenomenon as a compensatory mechanism that responds to increased excitation over the same preonset period.This model prediction is interesting because it relates to experimental results obtained with kainic acid known to generate seizures in the CA3region(Khalilov et al.,2002):

Journal of Clinical Neurophysiology?Volume22,Number5,October2005Interictal to Ictal Transition in TLE

using in vitro intact hippocampus,authors demonstrated that kainate is responsible for ictal activity but at the same time augments the excitatory input to interneurons and thus lead to an enhancement of inhibition.With the model identi?cation, we can go further and hypothesize that the increase of inhibition level is selective,i.e.,it mainly relates to slow dendritic inhibition.Indeed,in the model,the two types of interneurons(GABA A slow and fast kinetics)receive positive input from pyramidal cells.However,because the subset of GABA A,fast interneurons is inhibited by the subset GABA A,slow interneurons(Banks et al.,2000),the increase of excitation level at both inputs will mainly act on this latter subset that represents interneurons projecting to the dendritic region of pyramidal cells.The second signi?cant change regarding dendritic inhibition is a noticeable decrease at seizure onset when fast oscillations appear(ONSET period). This mechanism could be also related to the increase of excitation that takes place before seizure during the preonset period:it can be hypothesized that a point is reached where inhibitory interneurons cannot compensate anymore for the increased excitation and where provided inhibition collapses due to a“fatigue process.”This hypothesis relates to the concept of“fragile inhibition”proposed by Wu and Leung (2001)from a current source density analysis of the dentate gyrus in the kainic acid model of temporal lobe epilepsy.The authors showed that the relatively strong inhibition in the dentate gyrus of kainic acid-treated rats can breakdown readily in the presence of a low dose of GABA A receptor antagonist and concluded that the compensatory inhibition in the brain in animals and humans with temporal lobe seizures is vulnerable,although underlying mechanisms have not been clearly identi?ed.

Role of Fast Somatic Inhibition in the Generation of Fast Oscillations at

Seizure Onset

At seizure onset,faster oscillations(beta and low gamma bands)were observed on the EEG signal recorded from the hippocampus,as demonstrated by the computation of PSDs.Our identi?cation results show that they are asso-ciated to a constant or increasing value of model parameter FSI,in contrast to the dramatic decrease of parameter SDI. These results suggest that observed fast oscillations can be mainly related to the activity of inhibitory interneurons tar-geting pyramidal cells in their perisomatic region(with GABA A,fast kinetics).In a previous study,we already dem-onstrated that a fast feedback inhibitory loop is necessary for the model to generate EEG signals with gamma oscillations (Wendling et al.,2002).In the present study,results are obtained from direct automatic identi?cation procedure ap-plied on real intracerebral EEG signal and support our pre-vious results.This hypothesis of EEG fast oscillations related to fast inhibitory interneurons is also supported by several other studies mainly dealing with normal neuronal tissue. Penttonen et al.(1998)suggested that rapid discharges re?ect rhythmic postsynaptic potentials in hyperpolarized pyramidal cells brought about by rhythmically discharging somatic-projecting https://www.doczj.com/doc/2612704444.html,ing computational models of in-terconnected neurons and interneurons as well as experimen-tal data obtained from hippocampal slices from rats,White et al.(2000)also ended with the conclusion that GABA A,slow and GABA A,fast interneurons exist as two different popula-tions and that they are useful for generating theta and gamma rhythms(CA1region).More recently,similar?ndings were obtained from current source density analysis in the rat intact hippocampus during awake,attentive behavior(Hajos et al., 2004).The authors conclude that their results support the hypothesis of a synaptic feedback model of gamma oscilla-tions primarily involving pyramidal cells and perisomatic-projecting inhibitory interneurons.

Development of Seizure Activity

After the onset of seizure,a rhythmic activity(quasi-sinusoidal or spiking,generally in the theta band)is typically observed in recorded EEG signals.This activity corresponds to the progressive development of ictal activity that generally involves other mesial structures.In the present study dealing with a single neuronal population,hypotheses about interac-tions between populations distributed over distant brain struc-tures are not taken into account.However,regarding ictal activity in the hippocampus,we ended in the?ve patients with a situation where(1)the excitation level was high,(2) the slow inhibition level was equivalent or lower compared with interictal,and(3)the fast inhibition level was very low compared to interictal activity.From this observation,two remarks can be made.First,the transition from rapid dis-charge to rhythmic theta-like activity is explained,in the model,by the exhaustion of fast inhibitory processes.Second, seizure activity seems to correspond to a situation where inhibition level is globally reduced compared to excitation.

CONCLUSION

This computational modeling study offered the unique opportunity to relate electrophysiologic patterns typically recorded with clinical electrodes during the transition from interictal to ictal activity in the human hippocampus to ictogenesis mechanisms and to generate hypotheses about these mechanisms.Results demonstrated that the transition from interictal to ictal activity cannot be merely explained by an increase in excitation and a decrease in inhibition,but rather by a variety of complicated time-varying ensemble interactions between pyramidal cells and interneurons with slow and fast GABA A kinetics.In particular,the model predicts an increase of excitation during the preonset period (10to50seconds before seizure onset),an abrupt drop of dendritic inhibition at seizure onset when a rapid discharge (fast oscillations in the gamma band)and a crucial role of perisomatically projecting interneurons in observed EEG gamma and theta ictal activity.To us,these predictions could be tested in future experiments.In slices,electrophysiologic recordings combined with current source density analysis could be used to study the spatial origin of epileptic bursts (apical dendrites and soma-basal region).In epileptic pa-tients,during long-term video-EEG recording,stimulation

F.Wendling et al.Journal of Clinical Neurophysiology?Volume22,Number5,October2005

protocols based on a paired-pulse paradigm could be used to follow some physiologic parameters such as excitability.At present,the level of detail in the proposed model(neuronal population that best corresponds to the nature of our obser-vations,i.e.,EEG signals)does not allow us to specify hypotheses about cellular and molecular mechanisms in-volved in the above described phenomena.However,recent works in the?eld of computational modeling suggest that dynamic scale changes could be performed in such situations to relate macroscopic phenomena to microscopic ones(Vit-torini et al.,2004).Finally,as far as the clinical impact of this work is concerned,we think that the presented model-based approach is able to reveal physiologic changes that take place inside recorded brain structures during the transition from interictal to seizure activity.Future works will be oriented toward the development of time-optimized identi?cation pro-cedures to process long duration signals(over a sliding window).Indeed,such procedures could be used to detect excitation and inhibition changes typical of those occurring before seizure onset,and therefore,might be complementary to other methods already developed in the?eld of seizure prediction(see Lehnertz and Litt[2005]for review).

APPENDIX

Evolutionary Algorithms

Evolutionary algorithms are stochastic search tech-niques,inspired on the theories of evolution and natural selection,which can be used to?nd an optimal con?guration for a given system within speci?c constraints(Holland, 1975).In these algorithms,each“individual”of a“popula-tion”is characterized by a set of parameters(or chromo-some).An initial population is created,usually from a set of random chromosomes,and this population will“evolve,”improving its global performance,by means of an iterative process.During this process,each individual is evaluated by means of a“?tness”function,and a new generation is produced by applying mutation and crossover operators on selected individuals that present high“?tness”values,with probabilities p m and p c respectively.Convergence and robust-ness properties of EAs have been largely studied in the literature(Beasley et al.,1993;Goldberg,1989;Michalewicz, 1994).These properties depend on:(1)adequate individual coding,(2)proper de?nition of the?tness function,and(3) selection of appropriate genetic operators for crossover and mutation.

Individual Representation and Initial Population

Each individual represents an instance of the model and is characterized by the triplet P?[EXC,SDI,FSI]of the model.To reduce the search space,values for parameters EXC,SDI,and FSI were bounded to the physiologically plausible intervals[1–10],[1–50],and[1–20],respectively. These intervals,which have been de?ned after an exhaustive exploration of the parameter space(Wendling et al.2002),are employed by the EA during the construction of the initial population and the application of genetic operators.The initial population is constituted of randomly generated indi-viduals.The three parameter values of a given individual are independently generated from a uniform distribution de?ned under the corresponding feasibility interval.

Individual Evaluation

As previously stated,model output not only depends on the set of parameter values P but also on the realization n r of the input random noise.For given parameter vector P,dif-ferent realizations of this noise lead to different model out-puts and thus to different realizations of F S(P)(and conse-quently of?).In our case,at least two different strategies can been de?ned for individual evaluation:(1)one and only one realization n r is used for all individuals and for the whole identi?cation process,and(2)a new realization is created for each generation.We retained this second strategy. Selection Method

Once the error function has been evaluated for each individual,selection is carried out by means of the“roulette wheel”method,adapted for function minimization,in which the probability of selecting a given individual depends on the value of its error function,divided by the sum of all the error values of the population(Beasley et al.,1993).Only standard genetic operators,de?ned for real-valued chromosomes,have been used in this work:“uniform crossover,”which creates two new individuals(offspring)from two existing individuals (parents),by randomly copying each allele from one parent or the other,depending on a uniform random variable;and “Gaussian mutation,”which creates a new individual by randomly changing the value of one allele(selected ran-domly),based on a Gaussian distribution around the current value.This mutation operator respects the bounds de?ned for each allele by truncating the mutation value,if it is necessary.

A detailed description of these genetic operators can be found elsewhere(Michalewicz,1994).

REFERENCES

Bancaud J,Talairach J.(1973)[Methodology of stereo EEG exploration and surgical intervention in epilepsy.]Rev Otoneuroophtalmol45:315–28. Banks MI,Li TB,Pearce RA.(1998)The synaptic basis of GABAA,slow. J Neurosci18:1305–17.

Banks MI,White JA,Pearce RA.(2000)Interactions between distinct GABA(A)circuits in hippocampus.Neuron25:449–57.

Bartolomei F,Wendling F,Regis J,Gavaret M,Guye M,Chauvel P.(2004) Pre-ictal synchronicity in limbic networks of mesial temporal lobe epi-lepsy.Epilepsy Res61:89–104.

Beasley D,Bull DR,Martin R.(1993)An overview of genetic algorithms: part2.Research topics.University Comput15:170–181.

Bernard C,Marsden DP,Wheal HV.(2001)Changes in neuronal excitability and synaptic function in a chronic model of temporal lobe epilepsy. Neuroscience103:17–26.

Cossart R,Dinocourt C,Hirsch JC,Merchan-Perez A,De Felipe,J,Ben-Ari Y,Esclapez M,Bernard C.(2001)Dendritic but not somatic GABAergic inhibition is decreased in experimental epilepsy.Nat Neurosci4:52–62. Dalby NO,Mody I.(2001)The process of epileptogenesis:a pathophysio-logical approach.Curr Opin Neurol14:187–92.

Dichter MA.(1997)Basic mechanisms of epilepsy:targets for therapeutic intervention.Epilepsia9:S2–6.

Engel J Jr,Babb,TL,Crandall PH.(1989)Surgical treatment of epilepsy: opportunities for research into basic mechanisms of human brain function. Acta Neurochir Suppl(Wien)46:3–8.

Journal of Clinical Neurophysiology?Volume22,Number5,October2005Interictal to Ictal Transition in TLE

Freeman WJ.(1978)Models of the dynamics of neural populations.Elec-troencephalogr Clin Neurophysiol34(suppl):9–18.

Goldberg DE.(1989)Genetic algorithms in search,optimization and ma-chine learning.Boston:Addison-Wesley.

Gorter JA,van Vliet,EA,Aronica E,Lopes da Silva FH.(2002)Long-lasting increased excitability differs in dentate gyrus vs.CA1in freely moving chronic epileptic rats after electrically induced status epilepticus.Hip-pocampus12:311–24.

Hajos N,Palhalmi J,Mann EO,Nemeth B,Paulsen O,Freund TF.(2004) Spike timing of distinct types of GABAergic interneuron during hip-pocampal gamma oscillations in vitro.J Neurosci24:9127–37. Hernandez AI,Carrault G,Mora F,Bardou A.(2002)Model-based inter-pretation of cardiac beats by evolutionary algorithms:signal and model interaction.Artif Intell Med26:211–35.

Holland JH.(1975)Adaptation in natural and arti?cial systems.Boston: MIT Press.

Houser CR,Esclapez M.(1996)Vulnerability and plasticity of the GABA system in the pilocarpine model of spontaneous recurrent seizures.Epi-lepsy Res26:207–18.

Jansen BH,Rit VG.(1995)Electroencephalogram and visual evoked poten-tial generation in a mathematical model of coupled cortical columns.Biol Cybern73:357–66.

Jefferys JG.(2003)Models and mechanisms of experimental epilepsies. Epilepsia12:44–50.

Khalilov I,Hirsch J,Cossart R,Ben-Ari Y.(2002)Paradoxical anti-epileptic effects of a GluR5agonist of kainate receptors.J Neurophysiol88:523–7. Lehnertz K,Litt B.(2005)The First International Collaborative Workshop on Seizure Prediction:summary and data description.Clin Neurophysiol 116:493–505.

Lopes da Silva FH,Hoeks A,Smits H,Zetterberg LH.(1974)Model of brain rhythmic activity.The alpha-rhythm of the thalamus.Kybernetik15:27–37. Michalewicz Z.(1994)Genetic algorithms?data structures?evolution programs.New York:Springer-Verlag.

Miles R,Toth K,Gulyas AI,Hajos N,Freund TF.(1996)Differences between somatic and dendritic inhibition in the hippocampus.Neuron 16:815–23.

Pena-Reyes CA,Sipper M.(2000)Evolutionary computation in medicine:an overview.Artif Intell Med19:1–23.

Penttonen M,Kamondi A,Acsady L,Buzsaki G.(1998)Gamma frequency oscillation in the hippocampus of the rat:intracellular analysis in vivo.Eur J Neurosci10:718–28.

Schwartzkroin PA.(1994)Cellular electrophysiology of human epilepsy. Epilepsy Res17:185–92.

Spencer SS,Guimaraes P,Katz A,Kim J,Spencer D.(1992)Morphological patterns of seizures recorded intracranially.Epilepsia33:537–45.Suffczynski P,Kalitzin S,Lopes Da Silva FH.(2004)Dynamics of non-convulsive epileptic phenomena modeled by a bistable neuronal network. Neuroscience126:467–84.

Thomson AM,Radpour S.(1991)Excitatory connections between ca1 pyramidal cells revealed by spike triggered averaging in slices of rat hippocampus are partially NMDA receptor mediated.Eur J Neurosci 3:587–601.

Traub RD,Jefferys JG,Whittington MA.(1999a)Fast oscillations in cortical circuits.Cambridge:MIT Press.

Traub RD,Whittington MA,Buhl EH,Jefferys JG,Faulkner HJ.(1999b)On the mechanism of the gamma?beta frequency shift in neuronal oscilla-tions induced in rat hippocampal slices by tetanic stimulation.J Neurosci 19:1088–105.

Velasco AL,Wilson CL,Babb TL,Engel J Jr.(2000)Functional and anatomic correlates of two frequently observed temporal lobe seizure-onset patterns.Neural Plast7:49–63.

Vittorini V,Iacono M,Mazzocca NGF.(2004)The OsMoSys approach to multi-formalism modeling of systems.Software Syst Model3:68–81. Wendling F,Bartolomei F,Bellanger JJ,Chauvel P.(2002)Epileptic fast activity can be explained by a model of impaired GABAergic dendritic inhibition.Eur J Neurosci15:1499–508.

Wendling F,Bellanger JJ,Bartolomei F,Chauvel P.(2000)Relevance of nonlinear lumped-parameter models in the analysis of depth-EEG epileptic signals.Biol Cybern83:367–78.

White JA,Banks MI,Pearce RA,Kopell NJ.(2000)Networks of interneu-rons with fast and slow gamma-aminobutyric acid type A(GABAA) kinetics provide substrate for mixed gamma-theta rhythm.Proc Natl Acad Sci U S A97:8128–33.

Whittington MA,Traub RD,Faulkner HJ,Stanford IM,Jefferys JG.(1997) Recurrent excitatory postsynaptic potentials induced by synchronized fast cortical oscillations.Proc Natl Acad Sci U S A94:12198–203.

Wilson HR,Cowan JD.(1972)Excitatory and inhibitory interactions in localized populations of model neurons.Biophys J12:1–24.

Wittner L,Eross L,Czirjak S,Halasz P,Freund TF,Magloczky Z.(2005) Surviving CA1pyramidal cells receive intact perisomatic inhibitory input in the human epileptic hippocampus.Brain128(Pt1):138–52.

Wittner L,Magloczky Z,Borhegyi Z,Halasz P,Toth S,Eross L,Szabo Z, Freund TF.(2001)Preservation of perisomatic inhibitory input of granule cells in the epileptic human dentate gyrus.Neuroscience108:587–600. Wu K,Leung LS.(2001)Enhanced but fragile inhibition in the dentate gyrus in vivo in the kainic acid model of temporal lobe epilepsy:a study using current source density analysis.Neuroscience104:379–96.

Wu K,Leung LS.(2003)Increased dendritic excitability in hippocampal ca1 in vivo in the kainic acid model of temporal lobe epilepsy:a study using current source density analysis.Neuroscience116:599–616.

F.Wendling et al.Journal of Clinical Neurophysiology?Volume22,Number5,October2005

工业园区发展建设经验交流材料

工业园区发展建设经验交流材料 工业园区自被省政府列为重点支持园区以来，已经从规划启动阶段，步入形态开发阶段和目前的经济效益与社会效益相协调的快速发展阶段。 三年来，在省政府、县委县政府及各县直部门的大力支持帮助下，园区的区位优势得到很大的发挥，并且在农科院和县科技局的帮助下，引进了一批高新技术及产品，发展了一批外向型企业，同时形成两个产业园区，即电力电器产业园和健康产业园，园区内新兴的如义兴铁塔、日新光伏等科技企业日益壮大。园区的区域性经济得到很大提高。从税收的100多万提高到的1200多万元，工业产值实现5亿元，比建区前增长10倍。 今天，工业园区已经成为我县县域经济的重要经济增长点之一，不断集聚着高新技术产业，认真回顾与总结过去的实践，为园区今后的发展积累了结验，为未来的建设奠定了基础。 一、加强领导，强化服务与保障体系 建立园区之初，县委、县政府就将园区列为本地区发展重点，并明确指出要充分发挥这里的区位优势、经济实力和科技人才优势，把工业园区建设成为高标准、高科技、高效益的新兴经济区。同时，决定组建园区党组、管委会、开发

公司，形成党、政、企三位一体的组织领导机构。三年来，这个组织机构很好地适应了园区开发建设的需要，充分发挥出宏观调控、统一管理、协调服务的职能。 二、制定规划，提出重点建设目标 规划是建设的龙头。在市科委及有关部门的指导下，园区先后制定了－2020年发展规划、“九五”计划及发展规划。发展规划经县政府同意实施后，有力调控了生产力布局，明确提出重点建设目标。目前，集中各方面的力量建设起步区电力电器产业园和健康产业园，建成以电力电器为主的产业基地和健康产业基地。截止目前，入园企业已达21家。 在建项目5个，即将动工兴建项目9个。规划的实施不仅对生产力要素起到调控作用，也起到了很好的环境效果和生态效应。 三、促进科技、经济与金融结合 建区伊始，园区就提出科技、经济与金融要紧密结合的发展思路。实践证明，这一思路是园区取得规模效益的有效出路。 科技与经济结合尤为典型。园区积极与渤海大学、农科院、畜牧所、农业技术推广中心等科研院所建立长期稳定协作关系，这种科技与经济的有效结合使园区科技贡献份额由建区初期的54%，提高到目前的64%。 四、吸纳了一批科技队伍

系统集成公司简介范文

系统集成公司简介范文 系统集成商是指具备系统资质，能对行业用户实施系统集成的企业。下面是系统集成公司简介范文，欢迎参阅。 系统集成公司简介范文1 广州系统集成公司，专业为客户提供结构化布线系统、网络技术工程、程控交换机系统安装、监控安防系统、一卡通系统、音视频系统、机房建设等系统方案设计、施工及维护的服务。 “全面满足，不断超越，永创新高，打造行业领跑者形象”，公司一直秉承“以市场为导向、以客户为中心”的发展理念，以“团结、务实、拼搏、创新”为宗旨，不断苦练内功，随时为广大客户提供最优质的产品与服务。 系统集成公司长久以来一直努力的目标，就是协助客户建立最具竞争力的信息化系统，即协助客户去规划、建设和维护高性能的网络系统、可靠的网络安全建设、智能建筑系统等。并在业界树立了良好的口碑和有了很好的发展。如今，开建智能的服务网络覆盖多个地方并都设有办事机构。自建立以来，开建智能坚持的目标从不曾改变，凭借着其日益成熟的经营理念和专业水平，开建智能必将协助客户获取更强的竞争力。 专业而经验丰富的技术人力资源。开建智能的全体员工拥有专业的技术知识，并在大型系统、结构化网络系统、远程通讯、办公自动化、系统技术支持，和软件编写方面拥有丰富的经验。

系统集成公司简介范文2 中国电信集团系统集成有限公司成立于1996年，是中国电信集团公司的全资子公司。公司旨在为大客户提供ICT整体解决方案、为电信运营商提供应用软件开发和IT服务支撑、为中小企业客户提供综合信息化服务。 公司依托于中国电信全国垂直一体化的三级营销服务体系和运行维护体系，凭借中国电信丰富的网络资源、专业的电信及IT技术、优秀的技术团队、广泛的客户资源和行业知识，致力于为电信运营商、政府、金融、企业提供网络基础设施建设、网络升级及改造、网络管理服务、网络及设备代维服务、设备租赁、应用软件集成及开发、IT 服务支撑等“一站式”服务。 公司在为电信运营商、全国性大客户进行一系列大型网络建设和服务的过程中，归纳总结了一整套项目管理方法，形成了独特、完善的项目管理体系和实力强大的核心团队。公司通过了ISO9001(2000)质量管理体系认证。同时，还获得了信息产业部颁发的“计算机信息系统集成一级资质”和“通信信息网络系统集成甲级资质”，是国内第一家拥有“双一级”资质的系统集成企业。 公司将站在客户的角度思考客户的业务运营，通过对客户业务运营流程以及信息化需求的全面理解，为客户提供创新而适用的综合信息化解决方案和ICT支撑服务，提升客户价值，与客户共同成长。 系统集成公司简介范文3 联通系统集成有限公司是中国联通的全资子公司，注册资金亿元，

建设经验交流发言稿

建设经验交流发言稿 一、创新理念，科学谋划项目落地。每一个重大项目，从招商、立项、实施到竣工投产、达产达标，都是一个艰苦的系统工程，没有精益求精的精神和超前谋划的理念是难以实施好的。为此，我们把项目论证、立项、实施的过程当作凝聚全镇上下齐心协力谋划发展、建设小池的过程。一是注重规划强引导。规划是建设的龙头，只有科学的规划才能助推建设的发展。根据滨江新区总体规划，重新定位和调整产业发展布局，积极编制项目修建性规划。同时，坚持原则性与灵活性相结合，对规划强制性要求在保持总量不变前提下，尽量满足项目建设需要，最大限度支持项目建设。二是突出特色强支撑。围绕纺织服装、新型建材、农副产品加工、生物医药等特色产业、优势产业，积极建立完善产业项目库，着力策划一批投资密度强、 1 技术含量高、产业关联度大、经济效益好、资源消费低、环境污染少、人力资源得到充分发挥、税收贡献大的大项目、好项目。三是组建专班强责任。在全镇各线抽调业务骨干，集中优势兵力，组建了物流园建设指挥部、新自来水厂建设指挥部、九江二桥连接线建设指挥部等10个重点项目建设专班，专门负责项目的推介和争取工作。同时，实行镇党委领导包重点项目责任制，每个项目由一名党委领导具体负责，协调解决有关问题，为项目建设提供了坚强的组织保障。

二、借力发力，积极争取项目投资。始终坚持在发挥资源优势上做文章，在优化经济结构上下功夫，努力实现外资与资源的优势互补，互利双赢，共同发展。一方面突出招商引资。强化园区服务管理，完善招商优惠政策，抓好工业用地储备，全面提升临港产业园、中部商贸物流园的建设档次，招好商、招大商，逐步实现由“招商引资”向“招商选资”转变。全面启动万亩绿色无公害蔬菜生产基地、莲籽湖农业综合开发项目，引进了投资亿元、占地200亩的哈佛生物化工项目；投资 亿元、占地100亩的捷玛换热设备项目；投资亿、占地100亩的循环经济项目等亿元以上项目3个。一方面突出争资立项。牢牢把握投入的“黄金期”，发扬“三千”精神，不辞千言万语，不顾千山万水，不惜千方百计，只争朝夕上项目，全力以赴抓项目。根据国家投资导向和产业政策，加快跑省进京步伐，重点做好污水垃圾处理、水利基础设施、民政民生福利等方面项目的策划申报工作。加强与上级部门的联络沟通和服务跟进，实现污水处理厂、垃圾处理场、5000吨多功能码头、 2 安全饮水和事关小池今后五年发展的十六大重点工程等一批有利于打基础管长远、调结构促转型的大项目、好项目取得新突破。 三、强化服务，全力保障项目建设。一年来，我们积极

施工项目部经验交流

施工项目部经验交流 项目管理经验交流材料根据集团公司开展“项目管理效益年”活动的要求，我项目部在优化施工方案、细化过程控制、整合有利资源、拓展创效领域等各个方面推行精细化管理。项目综合管理水平得到提升。下面，就本项目的工程施工总体状况及推行精细化管理、加强项目管理等方面的主要做法汇报 如下。 一、工程施工总体状况我工区施工管段全长5、47km，主要工程数量有：特大桥5座，大桥1座、市政公路改造下穿框架桥1座，铁路框架桥、框架涵、盖板涵、圆管涵、倒虹吸等小型结构物24座，正线路基3、243km。今年四月份，我工区新增了东南半环的施工任务，成绵乐正线里程范围为D2K161+565～ D2K164+960，主要包括大桥1座，中桥3座，区间路基0、73km，站场路基2、66km和南站站场改造等施工任务。通过优化资源配置、科学组织施工，目前我单位共完成产值6、7亿元。在重难点及关键节点工程施工中，继xx年度取得了动车所、抬落道、门式墩三大战役的胜利后，xx年上半年，我们又在成洛大道改造工程、牛龙路三线桥悬灌梁施工、成绵乐正线框架桥施工、东南环线府河大桥临建等多项重点、难点、节点工程中取得了突破、塑造了亮点、树立了形象。在攻克难点、突破节点、加快进度的同时，高标准保证质量，精细化控制成本，严要求确保安全，全面

化追求效益。项目成本控制较好，集团公司成绵乐指挥部、四公司后方机关对我单位的成本管理工作给予了充分认可和高度评价；安全防范到位，无任何安全事故发生；在动车所、门式墩、悬灌梁、东南环线府河桥等重要形象工程中创出了质量亮点。并被建设单位授予xx年度成绵乐客运专线标准化项目部和标准化架子队荣誉称号。 二、推行精细管理，追求最大效益的主要做法。如何在保证正常施工的同时，提高项目创效水平，处理好进度与效益、质量、安全的矛盾，最终达到工期可控、安全平稳、质量优良、效益最大的目标，是工程项目管理的核心内容。针对这一核心，我单位时刻把完全成本管理作为项目管理的中心工作，始终抓住成本控制的主题，推行全方位、全过程、精细化、持续化的责任成本管理。并注重在“人本管理”上讲和谐，在“核算业务”上求严密，在“方案预控”上求深入，在“资源整合”上重互补，在“过程控制”中求精细，拓展成本控制领域，提高项目创效能力。总结我们的做法，主要体现在以下四点。 （一）树立“大团队”观念，依靠团队管项目。“思路决定出路，观念决定行动”。在本项工程建设中，我们工区项目部作为一个上受局指挥部直接领导，下对各架子队直接管控的“中间团队”，如何发挥我们工区这一层“承上启下”的作用，对上服从安排、抓好落实，对下积极引导、提高效率，全面深入细致的抓好成本管理，是我们推行精细化管理首要解决的问题。在这一

经验交流材料重大工程项目建设经验交流材料

三一文库（https://www.doczj.com/doc/2612704444.html,）/心得体会范文/经验交流材料重大工程项目建设经验交流材料 省建筑陶瓷产业基地（下称基地）项目和××省建筑陶瓷产业基地铁路专用线及其配套设施（下称专用线）项目分别于XX年4月5日和20XX年7月19日列入全省第十九次、第二十八次重大项目调度会。在省委、省政府以及省发改委的精心指导和大力支持下，在项目责任人的领导下，各有关部门通力协作，顽强拼搏，两个重大项目建设取得了较大成绩。到目前为止，建陶基地引进建陶生产企业及配套企业60家，合同资金为125.3亿元，到位资金41.34亿元，其中建陶企业30家，生产线192条，建成生产线35条，在建生产线25条，形成年产能3亿平米。基地铁路专用线项目获得铁道部行政许可，于20XX年3月动工建设。目前八景火车站扩建增设2股到发线已完工，专用线及货场征地工作顺利结束，5个标段都完成了招投标，3个已开工建设。对比上级要求，上述重大项目建设进度达到或超过了省里规定的要求。我们的主要经验交流是： 一、锻造精神，创新理念，凝聚重大项目工作动力每一个重大项目，从招商、立项、实施到竣工投产、达产达标，都是一个

艰苦的系统工程，没有强劲的动力和一股子精神是难以实施好的。为了搞好建陶基地及其跌路专用线等重大项目，我们把项目论证、立项、实施的过程当作凝聚全市上下特别是全体干部齐心协力建设重点项目、发展高安经济的过程。建陶基地上马前，我们在全体领导干部中组织了“中部要崛起，高安怎么办”和“高安实现工业化应从什么产业着手”的讨论，在明确建筑陶瓷是当时高安最现实、最擅长的产业选择后，又组织了包括在广东佛山建陶界工作的高安人士参加的“建陶基地最佳选址”的征求意见等 活动。这些活动有效激发了高安人特别是市乡两级干部积极投身基地建设的热情，并形成了“招引项目不怕冷落，审批项目不怕繁杂，实施项目不怕困难，服务项目不怕琐碎”的重点项目建设精神。在规划阶段，我们又广泛征求了各界的意见，经市委主要领导概括提炼为“建设中部最具影响力的建陶产区”的发展目标，“建设一个功能齐全的产业园区、一个建陶陶瓷生产中心、一个产品研发中心、一个三产服务中心、一批全国知名品牌集聚中心”的“五个一”发展思路和“做产业、成系统、可持续”的发展理念，这为项目的规划和建设明确了任务，指明了路径，指引了方向，提供了精神和思想保证。在为高安发展建功立业精神的感召下，各相关部门和干部发扬了愚公移山的精神，特别是基地管委会全体干部更是日夜奔忙在灰尘滚滚的工地，睡在村部大房间的通铺，常年加班加点，牺牲节假日，以至于一段时间大部分干部累倒病倒。新明珠集团总裁叶德林原有在别处落户的意向，但最终

综合系统集成解决方案

综合系统集成解决方 案 Revised on November 25, 2020

For personal use only in study and research; not for commercial use For personal use only in study and research; not for commercial use 系统综合集成 解 决 方 案 二○一四年三月七日

目录

系统综合集成解决方案 一、项目背景 现代飞速发展的信息技术，给通信系统信息化建设带来全新革命，正在深刻改变着系统工作、管理、服务、保障的各个环节。信息化发展到当前阶段，用户已经不仅仅满足于传统语音这一单一的通信方式，如何考虑充分利用既有信息资源，减少资源重复建设，融合语音、视频、数据的多媒体通信，实现“所见即所得、所见即可通、所见即可处”和“畅通无处不在”的目标，提升系统的感知能力、执法能力、处置能力、管理能力、服务能力，推动系统的工作模式、管理模式、服务模式的创新，成为系统的迫切需求。 二、当前系统现状 目前，通信系统已建了视频会议系统、视频监控系统、语音系统、无线超短波系统等各自独立、自成体系的通信系统。 (一)视频会议 系统各级部署有多个厂家的视频会议系统，还有一些软终端,这些系统有的基于协议，有的基于SIP协议。

(二)视频监控 系统大部分单位部署了视频监控系统，视频监控建设地点主要分布在各级的港口、码头、办公区等点位。因建设方式不统一，采用的监控管理平台也不统一，主要有海康威视、前卫视讯、大华和其它厂家等品牌，前端摄像机既有模拟摄像机，也有高清摄像机。 （三）语音系统 语音系统为主要使用的各类程控交换机、IP交换机，用于实现单位内部电话通话、以及与PSTN电话网络的互通。 (四)超短波系统 总公司和下属分公司及直属单位建立了全区或部分区域联网的超短波通信网，使用多个厂家、不同系列的超短波设备。 现有的视频会议系统、视频监控系统、各类语音系统、超短波系统等，使用的品牌多样，协议制式不统一，硬件设备跨代多。各系统为各自独立的信息孤岛，相互独立，互不能兼容互通，存在全网统一管理难实现，多协议制式难融合，多系统互通难达成等问题。

工程项目建设管理经验交流材料

工程项目建设管理经验交流材料 文章标题：工程项目建设管理经验交流材料 **公路**至**段工程项目建设，在省交通厅、省高速公路建设局及**市交通局的关怀领导下，在各有关部门的支持配合下，经所有参建单位的共同努力，目前，已完成工程总工作量的66。 一、工程项目概况 **公路**至**段工程项目，全长53.15公里。其中：主线全长50.24公里，采用一级半幅公路标准建设；路基宽度12米；支线全长2.91公里（镜泊乡支线2.56公里，水产养殖场支线0.35公里），采用一般三级公路标准建设，路基宽度8.5米。 全线路基土石方271万立方米，防护工程圬工体积10.6万立方米，原设计路面采用沥青混凝土，因沥青涨价及货源紧张，经省高速公路建设局组织专家论证，省交通厅于20xx年6月22日以黑交发〔20xx〕226号文件批复，将原沥青混凝土路面改为水泥混凝土路面。 全线新建大桥1,413.84米/3座，中桥323.04米/5座，小桥229.2米/8座，涵洞85道；全线设交叉道64处，其中与三级公路交叉12处，与大车道交叉52处。设收费站一处。

工程建设用地132.3公顷。其中：耕地19.2公顷，林地108.9公顷，其他用地4.2公顷。 建设资金及构成：本项目概算总金额为48,723万元，其构成为：1、交通部8,800万元，2、开行贷款6,270万元，3、地方筹资33,653万元（其中：波兰贷款2,850万美元，约23,000万元人民币，其余10,635万元为国内商业银行贷款）。 标段划分情况：本工程共分为14个标段。其中：路基标段6个，桥梁标段3个，路面标段2个，房建标段1个，机电标段1个，监理标段1个。 建设总工期为36个月，20xx年10月18日开工，计划20xx年9月30日完工交付使用。 基本建设程序，按公路工程建设项目规定，严格履行了逐级申报审批手续，批文按要求进行了立卷归档。 二、工程进展情况及目前存在的问题 （一）工程进展情况

重大工程项目建设经验交流材料_1

重大工程项目建设经验交流材料 省建造陶瓷产业基地（下称基地）项目和××省建造陶瓷产业基地铁路专用线及其配套设施（下称专用线）项目分别于XX年4月5日和2011年7月19日列入全省第十九次、第二十八次重大项目调度会。在省委、省政府以及省发改委的精心指导和大力支持下，在项目责任人的领导下，各有关部门通力协作，坚韧拼搏，两个重大项目建设取得了较大成绩。到目前为止，建陶基地引进建陶生产企业及配套企业60家，合同资金为125.3亿元，到位资金41.34亿元，其中建陶企业30家，生产线192条，建成生产线35条，在建生产线25条，形成年产能3亿平米。基地铁路专用线项目获得铁道部行政许可，于2011年3月动工建设。目前八景火车站扩建增设2股到发线已完工，专用线及货场征地工作顺利结束，5个标段都完成了招投标，3个已开工建设。对照上级要求，上述重大项目建设进度达到或超过了省里规定的要求。我们的要紧经验交流是： 一、锻造精神，创新理念，凝结重大项目工作动力>每一具重大项目，从招商、立项、实施到竣工投产、达产达标，基本上一具艰难的系统工程，没有强劲的动力和一股子精神是难以实施好的。为了搞好建陶基地及其跌路专用线等重大项目，我们把项目论证、立项、实施的过程当作凝结全市上下特别是全体干部齐心协力建设重点项目、进展高安经济的过程。建陶基地上马前，我们在全体领导干部中组织了“中部要崛起，高安如何办”和“高安实现工业化应从什么产业着手”的讨论，在明确建造陶瓷是当时高安最现实、最擅长的产业挑选后，又组织了包括在广东佛山建陶界工作的高安人士参加的“建陶基地最佳选址”的征求意见等活动。这些活动有效激发了高安人特别是市乡两级干部积极投身基地建设的热情，并形成了“招引项目别怕冷降，审批项目别怕繁杂，实施项目别怕困难，服务项目别怕琐碎”的重点项目建设精神。在规划时期，我们又广泛征求了各界的意见，经市委要紧领导概括提炼为“建设中部最具妨碍力的建陶产区”的进展目标，“建设一具功能齐全的产业园区、一具建陶陶瓷生产中心、一具产品研发中心、一具三产服务中心、一批全国知名品牌集聚中心”的“五个一”进展思路和“做产业、成系统、可持续”的进展理念，这为项目的规划和建设明确了任务，指明了路径，指引了方向，提供了精神和思想保证。在为高安进展建功立业精神的感召下，各相关部门和干部发扬了愚公移山的精神，特别是基地管委会全体干部更是日夜奔忙在灰尘滚滚的工地，睡在村部大房间的通铺，常年加班加点，牺牲节假日，以至于一段时刻大部分干部累倒病倒。新明珠集团总裁叶德林原有在别处降户的意向，但最终为高安领导干部凌晨为其项目选地的精神和35天平坦XX亩土地的速度所折服，先后做出签约投资12亿元和增资12亿元的决策。 二、高举龙头，系统配套，抢抓重大项目招商机遇 建陶基地既是一具单独的重大项目，也是一具由多个重大项目组成的母项目，其快速建设和成型，得益于做最具吸引力的基地品牌，招最有妨碍力的知名企业，搞最强服务能力的配套系统，并使三方面相互强化。在打造基地品牌方面，经过扎实有效的工作，获得了省发改委的立项批复和省经贸委的授牌，中国建材联合会的中国建造陶瓷产业基地授牌，中国建造卫生陶瓷协会和《陶城报》的陶瓷行业新锐榜2011、2011两个年度产区第一名。在招商方面，经过派出特意力量服务其资源和生产要素调查论证，顺利引进最具妨碍力的建陶航母新中源，带来了新明珠、欧雅、恒达、伟达、英皇卫浴洁具等一大批闻名品牌陶瓷企业。在配套方面，立脚服务大基地大项目，引进劳动力培训、产品展贸、现代物流、机械模具、包装印刷、科技研发等项目30多个，投资3800万元、占地60亩的中国建陶实训中心主体工程也已完工，投资5亿元的汇诚实业综合配套服务中心已建成一期。基地品牌的塑造促进了大项目的降户，大项目的降户促进了配套项目的建设，趋于完善的配套又反过来提升了基地品牌，促进了招商，形成了良性循环，从而牢牢抓住了新一轮建陶产业转移的历史性机遇。 三、超前预备，改进操作，提高重大项目建设效能

重大工程项目建设经验交流材料

重大工程项目建设经验交流材料 省建筑陶瓷产业基地（下称基地）项目和××省建筑陶瓷产业基地铁路专用线及其配套设施（下称专用线）项目分别于XX年4月5日和20XX年7月19日列入全省第十九次、第二十八次重大项目调度会。在省委、省政府以及省发改委的精心指导和大力支持下，在项目责任人的领导下，各有关部门通力协作，顽强拼搏，两个重大项目建设取得了较大成绩。到目前为止，建陶基地引进建陶生产企业及配套企业60家，合同资金为亿元，到位资金亿元，其中建陶企业30家，生产线192条，建成生产线35条，在建生产线25条，形成年产能3亿平米。基地铁路专用线项目获得铁道部行政许可，于20XX年3月动工建设。目前八景火车站扩建增设2股到发线已完工，专用线及货场征地工作顺利结束，5个标段都完成了招投标，3个已开工建设。对比上级要求，上述重大项目建设进度达到或超过了省里规定的要求。我们的主要经验交流是： 一、锻造精神，创新理念，凝聚重大项目工作动力>每一个重大项目，从招商、立项、实施到竣工投产、达产达标，都是一个艰苦的系统工程，没有强劲的动力和一股子精神是难以实施好的。为了搞好建陶基地及其跌路专用线等重大项目，我们把项目论证、立项、实施的过程当作凝聚全市上下

特别是全体干部齐心协力建设重点项目、发展高安经济的过程。建陶基地上马前，我们在全体领导干部中组织了“中部要崛起，高安怎么办”和“高安实现工业化应从什么产业着手”的讨论，在明确建筑陶瓷是当时高安最现实、最擅长的产业选择后，又组织了包括在广东佛山建陶界工作的高安人士参加的“建陶基地最佳选址”的征求意见等活动。这些活动有效激发了高安人特别是市乡两级干部积极投身基地建设的热情，并形成了“招引项目不怕冷落，审批项目不怕繁杂，实施项目不怕困难，服务项目不怕琐碎”的重点项目建设精神。在规划阶段，我们又广泛征求了各界的意见，经市委主要领导概括提炼为“建设中部最具影响力的建陶产区”的发展目标，“建设一个功能齐全的产业园区、一个建陶陶瓷生产中心、一个产品研发中心、一个三产服务中心、一批全国知名品牌集聚中心”的“五个一”发展思路和“做产业、成系统、可持续”的发展理念，这为项目的规划和建设明确了任务，指明了路径，指引了方向，提供了精神和思想保证。在为高安发展建功立业精神的感召下，各相关部门和干部发扬了愚公移山的精神，特别是基地管委会全体干部更是日夜奔忙在灰尘滚滚的工地，睡在村部大房间的通铺，常年加班加点，牺牲节假日，以至于一段时间大部分干部累倒病倒。新明珠集团总裁叶德林原有在别处落户的意向，但最终为高安领导干部凌晨为其项目选地的精神和35天平整XX亩土地

系统集成商开发技巧

系统集成商开发技巧

一、客户调查： 系统集成商的开发 1、系统集成商的定义：是指具备系统资质，能对行业用户实施系统集成的企业。系统集成分为:设备系统集成商（硬件系统集成商、弱电系统集成商）、应用系统集成商。 2、系统集成商的调查分析：系统集成商一般都有比较专业而稳定的渠道，比如：有的主做银行项目，有的主做公安有点主做教育，有的主做电力方面的项目，并且规模不同的系统集成商他的需求点也是不一样的，在系统集成商开发刚启动的时候要通过各种渠道如：行业协会、网站、媒体、行业刊物等收集目标系统集成商的名单，一段时间后就能通过系统集成商的介绍和相互了解而进行行业内部调查了； 调查的目的是为了：1、了解行业内的情况，更快将自己变成专业的业务员，2、更准确地寻找目标客户；每个地方的系统集成商都会很多，其中适合和我们合作的只有一部分，所以要在调查的结果中进行取舍； 确定目标客户的条件可以根据每个地方的特点而各异，可参考的标准有： （1）、公司业务规模：主营业务，年销量是多少、客户现状、客户名称、业务、规模、性质地址、邮编、网址。 （2）现有相关产品的使用情况：用途、品牌、数量、使用年限 （3）组织结构与采购相关的部门名称和人员构成（决策树） （4）部门之间的汇报和配合。各个部门在采购中的作用 （5）个人信息姓名、住址、联系方式、经历、爱好、年龄、家庭情况、子女喜欢的运动等（6）客户的工作范围、性格特点、客户内部的人员关系、竞争信息竞争对手在客户内的合作历史 （7）竞争对手销售人员的姓名和销售特点，产品优势和劣势 二、竞争品牌的调查： 着重了解：1、系统集成商现在采用的是什么品牌？2、该品牌的质量价格、结算周期、系统集成商对该品牌的欠款情况？3、系统集成商采用的是该品牌什么档次的品种？4、该品牌采用了什么样的公关方法？5、该品牌的产品市场知名度？6、双方合作的稳定性？等等。 尤其注意的是要敏感地发现近期系统集成商业务开发迅速的品牌并能分析他们的销售策略，为我们的销售方法提供参考；可以和系统集成商的人员或该品牌负责系统集成商开发的业务员沟通，多吸取优秀的业务经验 三、开发准备

项目管理经验交流会材料

水城矿业（集团）公司贵州新建业矿建米箩矿项目管理经验交流会材料 新建业公司米箩项目部 2011年12月28日

水城矿业（集团）公司贵州新建业矿建米箩矿项目管理经验交流会材料 尊敬的蔡总、田总；尊敬的各位领导、各位来宾 大家好！ 欢迎大家在百忙之中莅临米箩项目部检查指导工作，我谨代表米箩项目部全体员工向出席今天现场会的各位领导、各位来宾表示热烈的欢迎和诚挚的谢意！向长期以来对我们贵州新建业公司米箩项目部爱护有加、关心有加、支持有加、信任有加、帮助有加的各级组织、各位领导和各兄弟单位表示衷心的感谢！在此，还要特别感谢格目公司米箩矿筹备处的牵头，新建业公司的信任，把这么重要的会议在我项目部召开。在感到非常荣幸的同时，也感觉到今后米箩项目部肩上的担子有多重，责任有多大。但是，米箩项目部全体员工，早已做好思想准备，打好最后一个战役，坚决按时完成米箩矿的建设任务，如期交付格目底公司投入生产。现在，我受建业公司副总经理、米箩、马场项目部总指挥、米箩项目部经理余国勇同志的委托，向在座的各位领导、各位来宾汇报米箩项目部在管理上的一些经验和做法，如有不妥之处，敬请提出批评和指正。 一、工程概况 米箩矿设计生产能力120万吨/年，首采工作面投产，井巷工程量10883m(含措施井)，米箩矿主副井场区2010年6月开始施工，因工农关系影响，2010年6月至2010年12月,完成进尺161.5m。2011

年1月至2011年12月26日止,完成形象进尺4490m，成巷进尺4350m，2011年1月至5月完成进尺978.1m。 2011年6月1日至2011年12月26日施工3511m,平均月进尺501.7m，目前有掘进工作面六个，备用掘进工作面两个（副井片区一个、风井井片一个），施工头平均单进83.6米/月，在藉工作面平均单进62.7月，主斜井设计施工断面18.25m222度下山，最高单进达98月。 根据后余工程量，通过项目部全体员工的共同努力和业主方、监理方等的大力支持，争取按期或提前移交井巷工程。 二、目标锁定 按合同约定，在业主方规定的时间内，按期移交井巷工程。 三、重点分析实现目标要克服的重重困难 1、本矿属于煤层群，揭过煤频繁。 2、井筒设计倾角大，主斜井-220、副斜井-250、回风斜井-250—-170，全凭绞车配合箕斗提升运输,二期工程运输环节多，排矸、进料，运输非常困难。 3、压缩工期，提前投产，工期短，任务重。 4、二期工程设计断面小，不能平行作业，制约了安全生产。 5、定额单价与市场价相比实属偏低等。

信息系统集成

信息系统集成，就是通过结构化的综合布线系统和计算机网络技术，将各个分离的设备(如个人电脑)、功能和信息等集成到相互关联的、统一和协调的系统之中，使资源达到充分共享，实现集中、高效、便利的管理。系统集成采用功能集成、网络集成、软件界面集成等多种集成技术。系统集成实现的关键在于解决系统之间的互连和互操作性问题，它是一个多厂商、多协议和面向各种应用的体系结构，需要解决各类设备、子系统间的接口、协议、系统平台、应用软件等与子系统、建筑环境、施工配合、组织管理和人员配备相关的一切面向集成的问题。 系统组成 信息系统集成主要包括以下几个子系统的集成： (1)硬件集成 使用硬件设备将各个子系统连接起来，例如使用路由器连接广域网等。 (2)软件集成 软件集成要解决的问题是异构软件的相互接口。 (3)数据和信息集成 数据和信息集成建立在硬件集成和软件集成之上，是系统集成的核心，通常要解决的主要问题包括： ?合理规划数据和信息 ?减少数据冗余 ?更有效地实现信息共享 ?确保数据和信息的安全保密 (4)技术与管理集成 企业的核心问题是经济效益，使各部门协调一致地工作，做到市场销售、产品生产和管理的高效运转，是系统集成的重要内容。 (5)人与组织机构集成 系统集成的最高境界，提高每个人和每个组织机构的工作效率，通过系统集成来促进企业管理和提高管理效率。 系统特点

系统集成能够最大限度地提高系统的有机构成、系统的效率、系统的完整性、系统的灵活性等，简化系统的复杂性，并最终为企业提供一套切实可行的完整的解决方案。 系统集成的本质就是最优化的综合统筹设计，一个大型的综合计算机网络系统，系统集成包括计算机软件、硬件、操作系统技术、数据库技术、网络通讯技术等的集成，以及不同厂家产品选型，搭配的集成，系统集成所要达到的目标-整体性能最优，即所有部件和成分合在一起后不但能工作，而且全系统是低成本的、高效率的、性能匀称的、可扩充性和可维护的系统。 数据量高速膨胀、网络异构化程度加深，加剧了数据的破碎性。企业拼合失落的“数据碎片”需求的升温，让企业数据集成演变为一个独立的、跨越式的技术门类。 与我们熟知的“信息孤岛”相比，“数据孤岛”形容的是IT环境中最基本元素—数据的 离散状态。毋庸置疑，数据是企业最重要的信息资产，但是在现实环境中，企业数据往 往会因天然的业务分隔或行政分划，或者是IT应用复杂性的增加而驻留在不同的应用程序、数据库和遗留系统中。如何把不同来源、格式和质量的数据通过逻辑或物理的方法 集中起来，实现企业级数据的全面共享，进而为企业决策等高级应用提供支持，提升数 据资产的价值，是数据集成技术肩负的使命所在。 除了集中企业内部分散的业务数据之外，数据集成还能够在并购、分拆等商业行为 发生时充当IT边界变更后企业数据的“黏合剂”。对于现阶段频繁发生的企业并购行为而言，数据集成能够帮助这些企业在短期内实现应用整合。 以苏格兰的哈利法克银行和苏格兰银行合并为例，由于专注的金融服务领域具有明 确的互补性，两家银行的结合被业界广为称道。但是要合并两家银行信用卡业务系统中 的6亿份数据记录却是件异常艰苦的工作。为了实现信用卡业务数据合并的目标，两家 银行启动了持续性的集成操作。首先，哈利法克银行将其核心信用卡应用软件从First Data Resources平台迁移到EDS平台；然后，利用甲骨文和Business Objects建立一 个简单的报表数据库，并对3亿份记录进行了数据更新；最后，运用Informatica PowerMart套件所提供的ETL（抽取、转换和装载）功能建立新的数据库，并将剩余的 3亿份记录添加其中。整个过程仅历时四个月，数据集成工具所提供的自动化建模、多数据源导入特性使合并工程的效率大幅提升。 多维度技术视角 数据集成与BI（商业智能）、数据库、数据仓库技术存在紧密的联动关系，同时它也被认为是能够提升企业业务响应能力的关键技术之一。由于数据集成对业务敏捷化的 关键支持能力，使其正在成为促进不同应用间协同的基本保证。要给“数据集成”下一个定义并不容易，对于具有不同技术背景、不同知识体系的人而言，其概念的差异性明显。

乡镇建设美丽乡村典型经验交流材料

乡镇建设美丽乡村典型经验交流材料 凝心聚力，打造和谐美丽新 各位领导、同志们： 根据会议安排，现将我乡村美丽乡村建设情况做如下汇报： 一、基本情况 乡地处皖浙交界，与接壤，面积平方公里，辖村社区，人口万余人。美丽乡村建设试点村村位于我乡南部，是一个竹产业支撑型的新农村建设示范村，全村总面积平方公里，总人口3405人。近年来，我乡始终坚持以“ 争创省级新农村示范乡、打造最宜居住乡村”统揽全局，立足于乡域特色，认真谋划新农村建设总体思路，把握重点、强化措施、狠抓落实，扎实有序地在全乡各村开展了新农村建设。试点村村扎实推进美丽乡村建设，实现了“三新一带动”，即全村竹产业新发展，农民生活新提高，村容村貌新变化，美丽乡村建设取得初步成效。 二、主要做法 （一）成立组织，加强领导 一是乡党委、政府高度重视，加大扶持、指导和规范，成立了由乡党委书记、乡长任组长的美丽乡村建设领导组，分管领导任常务副组长，并驻村指导组织实施；二是从政府各部门抽调精干力量组建了美丽乡村建设办公室，充分调动各方面力量为美丽乡村建设提供服务；三是明确了村两委人员的工作职责，建立了村组户三级网络。现我乡已初步建立“政府指导，村组主导，全民参与”的美丽乡村建设发展机制。 （二）明晰思路，科学规划 一是按照县委、县政府的发展要求，结合村的村情，通过多次会议讨论和调查研究，并带领经充分论证后明确村美丽乡村建设的目标：“争创全省一流，打造4a旅游景区”，并把发展乡村旅游作为村的可持续发展目标来抓。二是按照规划先行的原则，依托村文化生态资源的优势，先后编制了三个层次的规划：首先聘请专家编制了高起点的《村美丽乡村建设总体规划》；尔后聘请专家编制了具体的景观节点开发、建设及保护性规划；围绕打造“三a”旅游景区目标，组织编制《村旅游发展建设总体规划》。同时，我们十分注重规划的权威性，严格按规划办事，将规划落到实处。 （三）完善机制，开拓创新

工程项目建设管理经验交流材料

工程项目建设管理经验交流材料 工程项目建设管理经验交流材料 文章标题：工程项目建设管理经验交流材料 **公路**至**段工程项目建设，在省交通厅、省高速公路建设局及**市交通局的关怀领导下，在各有关部门的支持配合下，经所有参建单位的共同努力，目前，已完成工程总工作量的66。 一、工程项目概况 **公路**至**段工程项目，全长53.15公里。其中：主线全长50.24公里，采用一级半幅公路标准建设；路基宽度12米；支线全长2.91公里（镜泊乡支线2.56公里，水产养殖场支线0.35公里），采用一般三级公路标准建设，路基宽度8.5米。 全线路基土石方271万立方米，防护工程圬工体积10.6万立方米，原设计路面采用沥青混凝土，因沥青涨价及货源紧张，经省高速公路建设局组织专家论证，省交通厅于20xx年6月22日以黑交发〔20xx〕226号文件批复，将原沥青混凝土路面改为水泥混凝土路面。 全线新建大桥1,413.84米/3座，中桥323.04米/5座，小桥229.2米/8座，涵洞85道；全线设交叉道64处，其中与三级公路交叉12处，与大车道交叉52处。设收费站一处。 工程建设用地132.3公顷。其中：耕地19.2公顷，林地108.9公顷，其他用地4.2公顷。 建设资金来源及构成：本项目概算总金额为48,723万元，其构成为：1、交通部8,800万元，2、开行贷款6,270万元，3、地方筹资33,653万元（其中：波兰贷款2,850万美元，约23,000万元人民币，其余10,635万元为国内商业银行贷款）。 标段划分情况：本工程共分为14个标段。其中：路基标段6个，桥梁标段3个，路面标段2个，房建标段1个，机电标段1个，监理标段1个。 建设总工期为36个月，20xx年10月18日开工，计划20xx年9月30日完工交付使用。 基本建设程序，按公路工程建设项目规定，严格履行了逐级申报审批手续，批文按要求进行了立卷归档。 二、工程进展情况及目前存在的问题 （一）工程进展情况 截止20xx年8月26日，全线路基土石方已基本完成（除因山体滑坡段K26 300～K28 285改线段外），路基工程经检测合格后，均已交给了路面施工单位进行了路面下基层施工；涵洞工程基本完成；中、小桥已完成桥面铺装50；红岭高架桥11号块已经浇筑完毕；复兴泡大桥95片梁板，已预制完成84片尚差11片，已架梁70片；路面下基层施工已完成58,300㎡，占总量的11；防护工程已完成总量的25。今年开工以来完成工作量5,380万元，累计完成工作量32,159万元，占概算投资的60。 （二）目前存在的问题 一是本项目地方筹集资金大部分是引进外资（韩国政府贷款），因出现变化改为引进波兰贷款，致使工程建设资金短缺，影响制约了工程进度。

财政性投资项目管理经验交流材料

财政性投资项目管理经验交流材料XX年底以来，xxxx县坚持把创新制度建设作为加强财政性投资项目管理的抓手，着力完善和认真落实科学、民主、高效的项目决策和组织实施程序，财政性投资行为进一步规范，财政资金使用效益进一步提升，公平公开公正的项 目投资环境逐步形成。截至XX年11月底，完成项目财政评审179个、招标核准46个、结算审计46个，三个环节共节减建设资金 1.2亿元，审减率达20%以上。 一、强化制度设计，操作程序更加规范严密。 针对过来项目管理制度不够完善，预算评审环节缺失、招标投标流于形式、工程量增减随意、项目资金被挤占挪用等现象时有发生的问题，县委、县政府在借鉴外地先进经验、广泛征求部门意见基础上，研究出台财政投资评审管理暂行办法、招标投标管理办法（试行）（以下简称“两个办法”），并专门设立县财政投资评审中心和县政府投资审计中心，调剂9个专业技术人员编制，加强项目关键环节管理。一是严把财政评审的前置关。把财政投资评审报告作为预算控制、工程招标、资金拨付、办理财务决算的依据，严格实行先评审、后核定预算，先评审、后招标采购，先评审、后办理拨款，先评审、后批复决算，先评审、后移交资产的“五先五后”制度。实施预算评审后，送审各类项目资金4.94亿元，审减资金4483万元，审减率达9.1%。二是严把招投标的公正关。对规模以上项目，一律实行招标方

案必须由发改部门核准、监察部门备案，评标和定标必须由专家委员会执行，招标过程必须由监察、发改、财政、审计和行政主管部门监督，项目合理定价必须由业主单位提出、县政府按规模分层级合理下浮的“四个必须”制度；对规模以下项目，业主单位自行研究招标方案一律实行决策过程必须有纪检监察人员在内的2/3以上单位班子成员参会，必须经80%以上参会人员同意，必须由县政府分管领导审批的“三个必须”制度。今年来，先后核准招投标项目46个，其中公开招标35个、邀请招标11个，公开招标投资额占总投资额的86.5%，同比增长69.1%。三是严把审计结算的免疫关。对所有财政性投资项目，严格实行必须办理竣工结算审计方可支付尾款和强化项目设计变更跟踪审计、工程量签证跟踪审计的“一个必须、两个跟踪审计”制度，较好地发挥了审计“免疫系统”功能作用。今年来，共对46个项目进行竣工结算审计，送审资金2.9亿元，审减资金2816万元，审减率达9.7%，特别是对南山学校项目的跟踪审计，节约建设成本近200万元。 二、坚持集体决策，利益约束更加动真碰硬。 将财政投资的专项资金、政府性基金、政府性融资、政府采购、外资外援、bt等项目全部纳入集体决策范畴，科学有效管控项目利益。一是锁定核心利益不含糊。财政评审过程中，评审中心会同业主和主管、监督单位，共同会商项目效益，重点核实项目工程量、单价定额等涉及核心利益事项。XX年，该环节仅调整工程量、核定单价定额就

项目部精细化管理经验材料(最新)

推行精细化管理全面提升项目管理——四公司成绵乐项目部项目管理经验交流材料 根据集团公司开展“项目管理效益年”活动的要求，我项目部在优化施工方案、细化过程控制、整合有利资源、拓展创效领域等各个方面推行精细化管理。项目综合管理水平得到提升。下面，就本项目的工程施工总体状况及推行精细化管理、加强项目管理等方面的主要做法汇报如下。 一、工程施工总体状况 我工区施工管段全长5.47km，主要工程数量有：特大桥5座，大桥1座、市政公路改造下穿框架桥1座，铁路框架桥、框架涵、盖板涵、圆管涵、倒虹吸等小型结构物24座，正线路基3.243km。今年四月份，我工区新增了东南半环的施工任务，成绵乐正线里程范围为D2K161+565～D2K164+960，主要包括大桥1座，中桥3座，区间路基0.73km，站场路基2.66km和南站站场改造等施工任务。 通过优化资源配置、科学组织施工，目前我单位共完成产值6.7亿元。在重难点及关键节点工程施工中，继2010年度取得了动车所、抬落道、门式墩三大战役的胜利后，2011年上半年，我们又在成洛大道改造工程、牛龙路三线桥悬灌梁施工、成绵乐正线框架桥施工、东南环线府河大桥临建等多项重点、难点、节点工程中取得了突破、塑造了亮点、树立了形象。 在攻克难点、突破节点、加快进度的同时，高标准保证质量，精细化控制成本，严要求确保安全，全面化追求效益。项目成本控制较好，集团公司成绵乐指挥部、四公司后方机关对我单位的成本管理工作给予了充分认可和高度评价；安全防范到位，无任何安全事故发生；在动车所、门式墩、悬灌梁、东南环线府河桥等重要形象工程中创出

了质量亮点。并被建设单位授予2010年度成绵乐客运专线标准化项目部和标准化架子队荣誉称号。 二、推行精细管理，追求最大效益的主要做法。 如何在保证正常施工的同时，提高项目创效水平，处理好进度与效益、质量、安全的矛盾，最终达到工期可控、安全平稳、质量优良、效益最大的目标，是工程项目管理的核心内容。针对这一核心，我单位时刻把完全成本管理作为项目管理的中心工作，始终抓住成本控制的主题，推行全方位、全过程、精细化、持续化的责任成本管理。并注重在“人本管理”上讲和谐，在“核算业务”上求严密，在“方案预控”上求深入，在“资源整合”上重互补，在“过程控制”中求精细，拓展成本控制领域，提高项目创效能力。总结我们的做法，主要体现在以下四点。 （一）树立“大团队”观念，依靠团队管项目。 “思路决定出路，观念决定行动”。在本项工程建设中，我们工区项目部作为一个上受局指挥部直接领导，下对各架子队直接管控的“中间团队”，如何发挥我们工区这一层“承上启下”的作用，对上服从安排、抓好落实，对下积极引导、提高效率，全面深入细致的抓好成本管理，是我们推行精细化管理首要解决的问题。在这一问题上我们树立了“大团队”观念，这个团队既包括局指挥部这一“上层团队”，也包括工区项目部这一“中间团队”，还包括现场架子队这一“一线团队”，整个团队的目标和利益是一致的，都是在保证安全、优质、高效的前提下实现效益最大化。基于这种团队观念，我们对局指挥部坚持“绝对服从但不依赖，完全配合处处主动”的工作态度，对劳动层架子队推行“本级化、人性化、服务化”的管理模式，对本级管理人员实行“人文关怀、岗位锻炼、考核奖励”的激励机制，确保了这个“大团队”的和谐、和睦、团结，充分发挥了整个团队的潜能。