node27.html
来自「an analysis software with souce code for」· HTML 代码 · 共 143 行
HTML
143 行
<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN">
<!--Converted with LaTeX2HTML 96.1-h (September 30, 1996) by Nikos Drakos (nikos@cbl.leeds.ac.uk), CBLU, University of Leeds -->
<HTML>
<HEAD>
<TITLE>Questions of interpretation</TITLE>
<META NAME="description" CONTENT="Questions of interpretation">
<META NAME="keywords" CONTENT="Surrogates">
<META NAME="resource-type" CONTENT="document">
<META NAME="distribution" CONTENT="global">
<LINK REL=STYLESHEET HREF="Surrogates.css" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/Surrogates.css">
</HEAD>
<BODY bgcolor=#ffffff LANG="EN" >
<A NAME="tex2html345" HREF="node28.html" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node28.html"><IMG WIDTH=37 HEIGHT=24 ALIGN=BOTTOM ALT="next" SRC="next_motif.gif" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/next_motif.gif"></A> <A NAME="tex2html343" HREF="Surrogates.html" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/Surrogates.html"><IMG WIDTH=26 HEIGHT=24 ALIGN=BOTTOM ALT="up" SRC="up_motif.gif" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/up_motif.gif"></A> <A NAME="tex2html337" HREF="node26.html" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node26.html"><IMG WIDTH=63 HEIGHT=24 ALIGN=BOTTOM ALT="previous" SRC="previous_motif.gif" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/previous_motif.gif"></A> <BR>
<B> Next:</B> <A NAME="tex2html346" HREF="node28.html" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node28.html">Non-dynamic nonlinearity</A>
<B>Up:</B> <A NAME="tex2html344" HREF="Surrogates.html" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/Surrogates.html">Surrogate time series</A>
<B> Previous:</B> <A NAME="tex2html338" HREF="node26.html" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node26.html">Spike trains</A>
<BR> <P>
<H1><A NAME="SECTION00070000000000000000">Questions of interpretation</A></H1>
<A NAME="secinterpret"> </A>
Having set up all the ingredients for a statistical hypothesis test of
nonlinearity, we may ask what we can learn from the outcome of such a test.
The formal answer is of course that we have, or have not, rejected a specific
hypothesis at a given level of significance. How interesting this information
is, however, depends on the null hypothesis we have chosen. The test is most
meaningful if the null hypothesis is plausible enough so that we are prepared
to believe it in the lack of evidence against it. If this is not the case, we
may be tempted to go beyond what is justified by the test in our
interpretation. Take as a simple example a recording of hormone concentration
in a human. We can test for the null hypothesis of a stationary Gaussian linear
random process by comparing the data to phase randomised Fourier surrogates.
Without any test, we know that the hypothesis cannot be true since hormone
concentration, unlike Gaussian variates, is strictly non-negative. If we
failed to reject the null hypothesis by a statistical argument, we will
therefore go ahead and reject it anyway by common sense, and the test was
pointless. If we did reject the null hypothesis by finding a coarse-grained
``dimension'' which is significantly lower in the data than in the surrogates,
the result formally does not give any new information but we might be tempted
to speculate on the possible interpretation of the ``nonlinearity'' detected.
<P>
This example is maybe too obvious, it was meant only to illustrate that the
hypothesis we test against is often not what we would actually accept to be
true. Other, less obvious and more common, examples include signals which are
known (or found by inspection) to be non-stationary (which is not covered by
most null hypotheses), or signals which are likely to be measured in a static
nonlinear, but non-invertible way. Before we discuss these two specific caveats
in some more detail, let us illustrate the delicacy of these questions with a
real data example.
<P>
<blockquote><A NAME="977"> </A><IMG WIDTH=360 HEIGHT=252 ALIGN=BOTTOM ALT="figure1094" SRC="img175.gif" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/img175.gif"><BR>
<STRONG>Figure:</STRONG> <A NAME="figseizure"> </A>Intracranial neuronal potential recording during
an epileptic seizure (upper) and a surrogate data set with the same linear
correlations and the same amplitude distribution (lower). The data was kindly
provided by K. Lehnertz and C. Elger.<BR>
</blockquote>
<P>
<P><blockquote><A NAME="979"> </A><IMG WIDTH=328 HEIGHT=148 ALIGN=BOTTOM ALT="figure1095" SRC="img176.gif" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/img176.gif"><BR>
<STRONG>Figure:</STRONG> <A NAME="figbarcode"> </A>
Surrogate data test for the data shown in Fig.<A HREF="node27.html#figseizure" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node27.html#figseizure">18</A>. Since the
prediction error is lower for the data (longer line) than for 99 surrogates
(shorter lines), the null hypothesis may be rejected at the 99% level of
significance. The error bar indicates the mean and standard deviation of the
statistic computed for the surrogates.<BR>
</blockquote><P>
<P>
Figure <A HREF="node27.html#figseizure" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node27.html#figseizure">18</A> shows as an intra-cranial recording of the neuronal
electric field during an epileptic seizure, together with one iteratively
generated surrogate data set [<A HREF="node36.html#surrowe" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node36.html#surrowe">30</A>] that has the same amplitude
distribution and the same linear correlations or frequency content as the data.
We have eliminated the end-point mismatch by truncating the series to 1875
samples. A test was scheduled at the 99% level of significance, using
nonlinear prediction errors (see Eq.(<A HREF="node4.html#eqerror" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node4.html#eqerror">5</A>), <I>m</I>=3, <IMG WIDTH=37 HEIGHT=12 ALIGN=BOTTOM ALT="tex2html_wrap_inline2416" SRC="img177.gif" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/img177.gif">,
<IMG WIDTH=47 HEIGHT=12 ALIGN=BOTTOM ALT="tex2html_wrap_inline2418" SRC="img178.gif" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/img178.gif">) as a discriminating statistics. The nonlinear correlations we
are looking for should enhance predictability and we can thus perform a
one-sided test for a significantly <EM>smaller</EM> error. In a test with one data
set and 99 surrogates, the likelihood that the data would yield the smallest
error by mere coincidence is exactly 1 in 100. Indeed, as can be seen in
Fig. <A HREF="node27.html#figbarcode" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node27.html#figbarcode">19</A>, the test just rejects the null hypothesis.
<P>
<P><blockquote><A NAME="981"> </A><IMG WIDTH=356 HEIGHT=201 ALIGN=BOTTOM ALT="figure1096" SRC="img179.gif" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/img179.gif"><BR>
<STRONG>Figure:</STRONG> <A NAME="figseizdel"> </A>
Left: Same data set as in Fig. <A HREF="node27.html#figseizure" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node27.html#figseizure">18</A>, rendered in time delay
coordinates. Right: A surrogate data set plotted in the same way.<BR>
</blockquote><P>
<P>
Unfortunately, the test itself does not give any guidance as to what kind of
nonlinearity is present and we have to face notoriously ill-defined questions
like what is the most <EM>natural</EM> interpretation. Similar spike-and-wave
dynamics as in the present example has been previously reported [<A HREF="node36.html#FEEG" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node36.html#FEEG">47</A>] as
chaotic, but these findings have been questioned [<A HREF="node36.html#TEEG" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node36.html#TEEG">48</A>].
Hernández and coworkers [<A HREF="node36.html#cuba" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node36.html#cuba">49</A>] have suggested a stochastic limit cycle
as a simple way of generating spike-and-wave-like dynamics.
<P>
If we represent the data in time delay coordinates -- which is what we would
usually do with chaotic systems -- the nonlinearity is reflected by the
``hole'' in the centre (left panel in Fig. <A HREF="node27.html#figseizdel" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node27.html#figseizdel">20</A>). A linear
stochastic process could equally well show oscillations, but its amplitude
would fluctuate in a different way, as we can see in the right panel of the
same figure for an iso-spectral surrogate. It is difficult to answer the
question if the nonlinearity could have been generated by a static mechanism
like the measurement process (beyond the invertible rescaling allowed by the
null hypothesis). Deterministic chaos in a narrower sense seems
rather unlikely if we regard the prediction errors shown in
Fig. <A HREF="node27.html#figbarcode" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node27.html#figbarcode">19</A>: Although significantly lower than that of the
surrogates, the absolute value of the nonlinear prediction error is still more
than 50% of the rms amplitude of the data (which had been rescaled to unit
variance). Not surprisingly, the correlation integral (not shown here) does not
show any proper scaling region either. Thus, all we can hand back to the
clinical researchers is a solid statistical result but the insight into what
process <EM>is</EM> generating the oscillations is limited.
<P>
A recent suggestion for surrogates for the validation of <EM>unstable periodic
orbits</EM> (UPOs) may serve as an example for the difficulty in interpreting
results for more fancy null hypothesis. Dolan and coworkers [<A HREF="node36.html#witt" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node36.html#witt">24</A>]
coarse-grain amplitude adjusted data in order to extract a transfer matrix that
can be iterated to yield typical realisations of a Markov chain.<A NAME="tex2html25" HREF="footnode.html#982" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/footnode.html#982"><IMG ALIGN=BOTTOM ALT="gif" SRC="foot_motif.gif" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/foot_motif.gif"></A>
The rationale there is to test if the finding of a certain number of UPOs could
be coincidental, that is, not generated by dynamical structure. Testing
against an order <I>D</I> Markov model removes dynamical structure beyond the
``attractor shape'' (AS) in <I>D</I>+1 dimensions. It is not clear to us what the
interpretation of such a test would be. In the case of a rejection, they would
infer a dynamical nature of the UPOs found. But that would most probably mean
that in some higher dimensional space, the dynamics could be successfully
approximated by a Markov chain acting on a sufficiently fine mesh. This is at
least true for finite dimensional dynamical systems. In other words, we cannot
see what sort of dynamical structure would generate UPOs but not show its
signature in some higher order Markov approximation.
<P>
<BR> <HR>
<UL><A NAME="CHILD_LINKS"> </A>
<LI> <A NAME="tex2html347" HREF="node28.html#SECTION00071000000000000000" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node28.html#SECTION00071000000000000000">Non-dynamic nonlinearity</A>
<LI> <A NAME="tex2html348" HREF="node29.html#SECTION00072000000000000000" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node29.html#SECTION00072000000000000000">Non-stationarity</A>
</UL>
<HR><A NAME="tex2html345" HREF="node28.html" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node28.html"><IMG WIDTH=37 HEIGHT=24 ALIGN=BOTTOM ALT="next" SRC="next_motif.gif" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/next_motif.gif"></A> <A NAME="tex2html343" HREF="Surrogates.html" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/Surrogates.html"><IMG WIDTH=26 HEIGHT=24 ALIGN=BOTTOM ALT="up" SRC="up_motif.gif" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/up_motif.gif"></A> <A NAME="tex2html337" HREF="node26.html" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node26.html"><IMG WIDTH=63 HEIGHT=24 ALIGN=BOTTOM ALT="previous" SRC="previous_motif.gif" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/previous_motif.gif"></A> <BR>
<B> Next:</B> <A NAME="tex2html346" HREF="node28.html" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node28.html">Non-dynamic nonlinearity</A>
<B>Up:</B> <A NAME="tex2html344" HREF="Surrogates.html" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/Surrogates.html">Surrogate time series</A>
<B> Previous:</B> <A NAME="tex2html338" HREF="node26.html" tppabs="http://www.mpipks-dresden.mpg.de/~tisean/TISEAN_2.0/docs/surropaper/node26.html">Spike trains</A>
<P><ADDRESS>
<I>Thomas Schreiber <BR>
Mon Aug 30 17:31:48 CEST 1999</I>
</ADDRESS>
</BODY>
</HTML>
⌨️ 快捷键说明
复制代码Ctrl + C
搜索代码Ctrl + F
全屏模式F11
增大字号Ctrl + =
减小字号Ctrl + -
显示快捷键?