This repository has been archived on 2021-05-17. You can view files and clone it, but cannot push or open issues or pull requests.
scientificComputing/programming/lecture/sta_stc.tex

394 lines
11 KiB
TeX

\documentclass{beamer}
\usepackage{xcolor}
\usepackage{listings}
\usepackage{pgf}
%\usepackage{pgf,pgfarrows,pgfnodes,pgfautomata,pgfheaps,pgfshade}
%\usepackage{multimedia}
\usepackage[english]{babel}
\usepackage{movie15}
\usepackage[latin1]{inputenc}
\usepackage{times}
\usepackage{amsmath}
\usepackage{bm}
\usepackage[T1]{fontenc}
\usepackage[scaled=.90]{helvet}
\usepackage{scalefnt}
\usepackage{tikz}
\usepackage{ textcomp }
\usepackage{soul}
\usepackage{hyperref}
\definecolor{lightblue}{rgb}{.7,.7,1.}
\definecolor{mygreen}{rgb}{0,1.,0}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\mode<presentation>
{
\usetheme{Singapore}
\setbeamercovered{opaque}
\usecolortheme{tuebingen}
\setbeamertemplate{navigation symbols}{}
\usefonttheme{default}
\useoutertheme{infolines}
% \useoutertheme{miniframes}
}
\AtBeginSection[]
{
\begin{frame}<beamer>
\begin{center}
\Huge \insertsectionhead
\end{center}
% \frametitle{\insertsectionhead}
% \tableofcontents[currentsection,hideothersubsections]
\end{frame}
}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%5
\setbeamertemplate{blocks}[rounded][shadow=true]
\title[]{Introduction to Scientific Computing -- \\
Cross-Correlation, Spike--Triggered--Average and Reverse Reconstruction}
\author[]{Jan Grewe\\Abteilung f\"ur Neuroethologie\\
Universit\"at T\"ubingen}
\institute[Introduction to Scientific Computing]{}
\date{2014/10/13 - 2014/11/07}
%\logo{\pgfuseimage{../../resources/UT_BM_Rot_RGB.pdf}}
\subject{Einf\"uhrung in wissenschaftliche Datenverarbeitung}
\vspace{1em}
\titlegraphic{
\includegraphics[width=0.5\linewidth]{../../resources/UT_WBMW_Rot_RGB}
}
%%%%%%%%%% configuration for code
\lstset{
basicstyle=\ttfamily,
numbers=left,
showstringspaces=false,
language=Matlab,
commentstyle=\itshape\color{darkgray},
keywordstyle=\color{blue},
stringstyle=\color{green},
backgroundcolor=\color{blue!10},
breaklines=true,
breakautoindent=true,
columns=flexible,
frame=single,
captionpos=b,
xleftmargin=1em,
xrightmargin=1em,
aboveskip=10pt
}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\newcommand{\mycite}[1]{
\begin{flushright}
\tiny \color{black!80} #1
\end{flushright}
}
\input{../../latex/environments.tex}
\makeatother
\begin{document}
\begin{frame}[plain]
\frametitle{}
\vspace{-1cm}
\titlepage % erzeugt Titelseite
\end{frame}
\begin{frame}
\frametitle{Introduction to scientific computing}
\frametitle{Menue}
\begin{enumerate}
\item Cross-correlation
\item Estimation of the Spike--Triggered--Average -- STA
\item Reverse reconstruction
\end{enumerate}
\end{frame}
\begin{frame}[plain]
\huge{1. Recapitulation: Plotting neuronal activity as a function of time.}
\end{frame}
\begin{frame}
\frametitle{Recapitulation}
\only<1>{
\framesubtitle{Displaying the neuronal response over time - Rasterplot}
\begin{figure}
\centering
\includegraphics[width=0.375\columnwidth]{images/rasterplot}
\end{figure}
}
\only<2>{
\framesubtitle{Displaying the neuronal response over time - PSTH}
\begin{figure}
\centering
\includegraphics[width=0.5\columnwidth]{images/conv}
\end{figure}
}
\end{frame}
\begin{frame}
\frametitle{Relating stimulus and response}
\framesubtitle{Displaying the neuronal response over time - PSTH}
\begin{itemize}
\item What does this tell us? \pause
\item Wouldn't it be more interesting to relate the response to the stimulus?!
\end{itemize}
\end{frame}
\begin{frame}[plain]
\huge{2. Relating stimulus and response}
\end{frame}
\begin{frame}
\frametitle{Relating stimulus and response}
\framesubtitle{How can we relate the response to the stimulus?}
\begin{figure}
\centering
\includegraphics[height=0.75\textheight]{images/conv_stim}
\end{figure}
\end{frame}
\begin{frame}
\frametitle{Relating stimulus and response}
\framesubtitle{Cross--correlation}
The cross - correlation for (time) discrete data is defined as:
\begin{equation}
Q_{rs}(\tau) = \frac{1}{2\cdot N-|\tau|} \displaystyle\sum^{T}_{\tau=-T}{r(t) \cdot s(t-\tau)} \cdot \Delta \tau
\end{equation}
\pause
Don't worry, this is already implemented in MATLAB!
\end{frame}
\begin{frame}[fragile]
\frametitle{Relating stimulus and response}
\framesubtitle{Cross--correlation}
\footnotesize
\begin{lstlisting}
[c,lags] = xcorr(stimulus, response, 10000, 'unbiased');
fig = figure();
set(fig, 'PaperUnits', 'centimeters');
set(fig, 'PaperSize', [11.7 9.0]);
set(fig, 'PaperPosition',[0.0 0.0 11.7 9.0]);
set(fig,'Color', 'white')
plot(lags/sample_rate, c)
xlabel('lag [s]')
ylabel('correlation')
\end{lstlisting}
\end{frame}
\begin{frame}
\frametitle{Relating stimulus and response}
\framesubtitle{Cross--correlation}
\begin{figure}
\includegraphics[width=0.5\linewidth]{images/correlation}
\end{figure}
\end{frame}
\begin{frame}[fragile]
\frametitle{Relating stimulus and response}
\framesubtitle{Cross--correlation - Exercises}
\begin{enumerate}
\item Calculate the cross-correlation between two vectors of random
numbers.
\item Calculate the cross-correlation between one of these vectors
and itself (auto-correlation).
\item Calculate the cross-correlation between one vector and a
time-shifted version of itself (use \verb+circshift+ to do this).
\item Generate two vectors of random numbers, one having a (slight)
correlation with the other.
\item Calculate the correlation coefficient (\verb+corrcoef+).
\item Calculate the cross-correlation.
\item Calculate the the correlation coefficient between the one
vector and a \verb+circshif+ted version of the other.
\item Calculate the cross-correlation of these.
\item Find out the maximum correlation and its position.
\end{enumerate}
\textbf{Note:} Select max\_lag to be less than 10\% of the length of
your vectors!
\end{frame}
\begin{frame}[fagile]
\frametitle{Relating stimulus and response}
\framesubtitle{Cross--correlation - Exercises}
\begin{enumerate}
\item Create the cross correlation of the p-unit data and stimulus.
\item \textbf{Note:} you have to convert the spike\_times to a PSTH!
\item Find out the position of the correlation peak.
\item What does it tell you?
\end{enumerate}
\end{frame}
\begin{frame}[plain]
\huge{2. Spike--Triggered--Average --- STA}
\end{frame}
\begin{frame}[fragile]
\frametitle{Relating stimulus and response}
\framesubtitle{Spike--Triggered--Average --- STA}
The \textbf{STA} is the average stimulus that led to a spike in the
neuron.
\begin{equation}
STA(\tau) = \left\langle \frac{1}{n} \displaystyle\sum^{i=1}_{n}{s(t_i - \tau)} \right\rangle
\end{equation}
Sadly, we need to implement this ourselves.\newline\newline\pause
\vspace{1em}
\noindent
\textbf{Algorithm:}
\begin{enumerate}
\item For each spike a snippet is taken from the respective stimulus.
\item The snippets are averaged.
\end{enumerate}
\end{frame}
\begin{frame}[fragile]
\frametitle{Relating stimulus and response}
\framesubtitle{Spike--Triggered--Average -- STA}
\vspace{-1em}
\begin{figure}
\centering
\includegraphics[width=0.65\columnwidth]{images/sta}
\end{figure}
\pause
\vspace{-0.5em}
What does the \textbf{STA} tell us?
\begin{enumerate}
\item Is there a relation between stimulus and response? \pause
\item Is there a lag between them and how large is it? \pause
\item How far in the past does a neuron encode? \pause
\item Can it see into the future?
\end{enumerate}
\end{frame}
\begin{frame}[fragile]
\frametitle{Relating stimulus and response}
\framesubtitle{STA -- Exercises}
\textbf{Exercise:}
\begin{enumerate}
\item Write a function \verb+sta(x, y, count, sample_rate)+ that
takes the stimulus (x), the response (y, as spike times), the
number (count) of sample points it should cut out from the
stimulus and the sample\_rate to convert from times to
indices.
\item \textbf{Beware:} sometimes the spike\_time may be too close
to the beginning or the end of the stimulus to cut out enough
data.
\item Calculate the STA for P-unt data.
\end{enumerate}
\end{frame}
\begin{frame}[plain]
\huge{3. Reverse reconstruction using the \textbf{STA}}
\end{frame}
\begin{frame}[fragile]
\frametitle{Relating stimulus and response}
\framesubtitle{Reverse reconstruction using the \textbf{STA}}
\textbf{Basic idea:}\newline
\begin{itemize}
\item The \textbf{STA} is the average stimlus that led to a spike. \pause
\item The other way round: Whenever we observe a spike, the stimulus
was likely to have looked like the \textbf{STA}.\pause
\item Thus, we should be able to reconstruct how the stimulus looked
like to evoke the observed response.
\end{itemize}
\end{frame}
\begin{frame}[fragile]
\frametitle{Relating stimulus and response}
\framesubtitle{Reverse reconstruction using the \textbf{STA}}
\textbf{Algorithm:}\newline
\begin{enumerate}
\item Estimate the \textbf{STA}. \pause
\item Create a binary representation of the spike train (vector of
\verb+zeros+ in which each \textbf{1} signals the occurence of a
spike).\pause
\item Use the STA as a Kernel (compare to the spike convolution
method to create the PSTH) and convolve (\verb+conv+) it with the
\textbf{STA}.
\end{enumerate}
\end{frame}
\begin{frame}
\frametitle{Relating stimulus and response}
\framesubtitle{Reverse reconstruction using the \textbf{STA}}
\textbf{Exercise:}
\begin{enumerate}
\item Reconstruct the stimulus from the P-unit data.
\item Plot original and reconstructed stimulus into the same plot.
\item Are they similar? In which instances are they different?
\item Estimate the differences in terms of the
\textit{mean-square-error} you have used in week 2. Normalize it
to the standard-deviation of the original stimulus.
\item Estimate the \textbf{STA} on the basis of the first half of
the data and reconstruct the stimulus of the second half.
\item Compare the reconstruction errors.
\item How strongly does the error depend on the amount of data you
use?
\end{enumerate}
\end{frame}
\end{document}
\begin{frame}[plain]
\huge{4. Spike--Triggered--Covariance \textbf{STC}}
\end{frame}
\begin{frame}
\frametitle{Relating stimulus and response}
\framesubtitle{Spike--Triggered--Covariance \textbf{STC}}
\textbf{Problem:}
\begin{itemize}
\item The \textbf{STA} is a linear filter.
\item It can only reveal linear relationships between stimulus and response.\pause
\end{itemize}
\vspace{1em}
Further relations can be recovered using the Spike-triggered-covariance.\pause
\begin{equation}
STC = \frac{1}{N-1} \cdot \displaystyle\sum_{n=1}^{N}[\overrightarrow{s}(t_n) - STA]
[\overrightarrow{s}(t_n) - STA]^T
\end{equation}
where:\\ $N$ is the total number of spikes,
$\overrightarrow{s}(t_n)$ the stimulus snippet for the $n^{th}$
spike, and $STA$ the Spike-Triggered-Average.
\end{frame}
\begin{frame}
\frametitle{Relating stimulus and response}
\framesubtitle{Spike--Triggered--Covariance}
\begin{itemize}
\item
\end{itemize}
\end{frame}
\end{document}