How to construct qqplot confidence bands

Dear forum users,

I need to construct 95% white noise confidence intervals in the qqplot figure. Can anyone help me how to construct these. I am using the code

qqplot(y)

I want to test the hypothesis that "y" is a vector of white noise. Can i just take the line computed by qqplot and add/subtract 1.96*sqrt(length(y))??

[Jonas]

One way to obtain confidence intervals is to use simulations

    qqplot(...)
        
    %%
    % Small sample size => use simulations:
    nSim = 3000;     % use this many data-point to simulate the quantiles
    confidenceInterval = 0.95;
    alpha = 1-confidenceInterval; % alpha risk
    n     = length(data);
    xSim  = sort(randn(nSim,n),2);           % sorted random variables
    zSim  = mean(data) + xSim*std(data);     % quantiles N(mu, sd)
    qSim = quantile(zSim, [alpha/2, 1-alpha/2]);  % quantiles
    
    hold on
    xVals = qnorm(([1:n]-0.5)/n);            % theoretical quantiles
    plot(xVals, qSim(:,:), 'b--');             
    hold off

[Morten Nissov]

I don't think qnorm is a matlab function?

[Scott MacKenzie]

In File Exchange there is a qnorm function