Problem with measuring the uncertainty of the fit parameters

Hello, I am trying to fit 7 data points with errorbars in y axis to the equation y=mx.

I am using fit function and choosing the confidence bound to be 68% using 'confint' function to find the value of standard deviation σ from it . Additionally, when I am measuring the uncertainty by finding the place where minimum going to , I get a different value for σ. I have also measured σ manually by using equations of least squared method in pen and papper. The value of this σ is same as the σ that I found with minimum method. I want to know what is the reason behind this discrepancy? Am I doing something wrong?

[linE_y,gfl_y] = fit((10:10:70)',AVG_MSD_tracks(2:8),ftlin,'Weight',WeightMSD(2:8));
parameters_linE =coeffvalues(linE_y);
error_parameters_linE=confint(linE_y,0.68);
t_linE=tinv((1+0.68)/2,gfl_y.dfe);
error_Dv_linE= abs((error_parameters_linE(1,1)-error_parameters_linE(2,1)))/(2*t_linE);

NOTE: I have tried using origin linear fit and the σ is same as the σ found using MATLAB.