FFT. It doesn't seem to find the frequencies

Hi, Im having a lot of issues turning my time domain into a frequency domain. What am I doing wrong? I expect some high frequency noise - also outside of 500 Hz, and the fft only shows frequencies around 0 Hz...

%% Load data
load('henrikrigtigforsoeg.mat')                             
data1 = data(1:20000)*0.0049;                     
fs  = 1000;                                                 % Sampling frequency
time = 1/fs;                                       
leng = 20000;                                           %length of signal
t = (0:leng-1)*time                                     % Time vector

%% Cut the data
kont_1 = data1(3000:7000);                          %Contraction 1
t_y1 = (0:4000)*time;                                    %time vector for contraction 1
leng_k1 = length(kont_1);                            %length of the contraction

%% FFT
Y1 = fft(kont_1);                                      % Lav en FFT på data
P2_y1 = abs(Y1/leng_k1);
P1_y1 = P2_y1(1:leng_k1/2+1);
P1_y1(2:end-1) = 2*P1_y1(2:end-1);
f_y1 = fs*(0:(leng_k1/2))/leng_k1;                         

%% Plot data
figure
subplot(2,1,1)                                          % Plot the raw data
plot(t_y1,kont_1)
title('Contraction in time-domain')
xlabel('Time in seconds')
ylabel('Magnitude in V')

subplot(2,1,2)                                         % Plot the Fourier transformation
plot(f_y1,P1_y1)
title('Single-sided power spectrum of first contraction')
xlabel('Frequencies in Hz')
ylabel('Power in V')

[Adam]

A high frequency component at 0 means your data is very badly zero-centred. If you subtract the mean from your data this should go away.

[Maria Goncalves Møller]

Like the following right?

data1 = (data(1:20000)*0.0049);

data2 = data1- mean(data1)

It still looks like this: