Example #1
0
relevant_features = np.random.randint(0, n_features, 10)
for i in relevant_features:
    w[i] = stats.norm.rvs(loc=0, scale=1. / np.sqrt(lambda_))
# Create noite with a precision alpha of 50.
alpha_ = 50.
noise =  stats.norm.rvs(loc=0, scale=1. / np.sqrt(alpha_), size=n_samples)
# Create the target
y = np.dot(X, w) + noise

################################################################################
# Fit the ARD Regression
clf = ARDRegression(compute_score = True)
clf.fit(X, y)

ols = LinearRegression()
ols.fit(X, y)

################################################################################
# Plot the true weights, the estimated weights and the histogram of the
# weights
pl.figure(figsize=(6, 5))
pl.title("Weights of the model")
pl.plot(clf.coef_, 'b-', label="ARD estimate")
pl.plot(ols.coef_, 'r--', label="OLS estimate")
pl.plot(w, 'g-', label="Ground truth")
pl.xlabel("Features")
pl.ylabel("Values of the weights")
pl.legend(loc=1)

pl.figure(figsize=(6, 5))
pl.title("Histogram of the weights")
Example #2
0
def test_contrast(p_intra, p_inter, num_trials, data_path, muscimol_amount=0*nS, injection_site=0, single_inh_pop=False):
    num_groups=2
    trial_duration=1.0*second

    wta_params=default_params()
    wta_params.p_b_e=0.1
    wta_params.p_x_e=0.1
    wta_params.p_e_e=p_intra
    wta_params.p_e_i=p_inter
    wta_params.p_i_i=p_intra
    wta_params.p_i_e=p_inter
    input_sum=40.0

    contrast_range=[0.0, 0.0625, 0.125, 0.25, 0.5, 1.0]
    trial_contrast=np.zeros([len(contrast_range)*num_trials,1])
    trial_max_bold=np.zeros(len(contrast_range)*num_trials)
    trial_max_exc_bold=np.zeros(len(contrast_range)*num_trials)
    for i,contrast in enumerate(contrast_range):
        print('Testing contrast %0.4f' % contrast)
        inputs=np.zeros(2)
        inputs[0]=(input_sum*(contrast+1.0)/2.0)
        inputs[1]=input_sum-inputs[0]

        for j in range(num_trials):
            print('Trial %d' % j)
            trial_contrast[i*num_trials+j]=contrast
            np.random.shuffle(inputs)

            input_freq=np.zeros(num_groups)
            for k in range(num_groups):
                input_freq[k]=float(inputs[k])*Hz

            file='wta.groups.%d.duration.%0.3f.p_b_e.%0.3f.p_x_e.%0.3f.p_e_e.%0.3f.p_e_i.%0.3f.p_i_i.%0.3f.p_i_e.%0.3f.contrast.%0.4f.trial.%d.h5' %\
                 (num_groups, trial_duration, wta_params.p_b_e, wta_params.p_x_e, wta_params.p_e_e, wta_params.p_e_i,
                  wta_params.p_i_i, wta_params.p_i_e, contrast, j)

            out_file=None
            if data_path is not None:
                out_file=os.path.join(data_path,file)
            wta_monitor=run_wta(wta_params, num_groups, input_freq, trial_duration, record_neuron_state=True,
                output_file=out_file, muscimol_amount=muscimol_amount, injection_site=injection_site, single_inh_pop=single_inh_pop)

            trial_max_bold[i*num_trials+j]=np.max(wta_monitor.voxel_monitor['y'].values)
            trial_max_exc_bold[i*num_trials+j]=np.max(wta_monitor.voxel_exc_monitor['y'].values)

    x_min=np.min(contrast_range)
    x_max=np.max(contrast_range)

    fig=plt.figure()
    clf=LinearRegression()
    clf.fit(trial_contrast,trial_max_bold)
    a=clf.coef_[0]
    b=clf.intercept_

    plt.plot(trial_contrast, trial_max_bold, 'x')
    plt.plot([x_min,x_max],[a*x_min+b,a*x_max+b],'--')
    plt.xlabel('Input Contrast')
    plt.ylabel('Max BOLD')
    plt.show()

    fig=plt.figure()
    clf=LinearRegression()
    clf.fit(trial_contrast,trial_max_exc_bold)
    a=clf.coef_[0]
    b=clf.intercept_

    plt.plot(trial_contrast, trial_max_exc_bold, 'o')
    plt.plot([x_min,x_max],[a*x_min+b,a*x_max+b],'--')
    plt.xlabel('Input Contrast')
    plt.ylabel('Max BOLD (exc only)')
    plt.show()
Example #3
0
def test_contrast(p_intra,
                  p_inter,
                  num_trials,
                  data_path,
                  muscimol_amount=0 * nS,
                  injection_site=0,
                  single_inh_pop=False):
    num_groups = 2
    trial_duration = 1.0 * second

    wta_params = default_params()
    wta_params.p_b_e = 0.1
    wta_params.p_x_e = 0.1
    wta_params.p_e_e = p_intra
    wta_params.p_e_i = p_inter
    wta_params.p_i_i = p_intra
    wta_params.p_i_e = p_inter
    input_sum = 40.0

    contrast_range = [0.0, 0.0625, 0.125, 0.25, 0.5, 1.0]
    trial_contrast = np.zeros([len(contrast_range) * num_trials, 1])
    trial_max_bold = np.zeros(len(contrast_range) * num_trials)
    trial_max_exc_bold = np.zeros(len(contrast_range) * num_trials)
    for i, contrast in enumerate(contrast_range):
        print('Testing contrast %0.4f' % contrast)
        inputs = np.zeros(2)
        inputs[0] = (input_sum * (contrast + 1.0) / 2.0)
        inputs[1] = input_sum - inputs[0]

        for j in range(num_trials):
            print('Trial %d' % j)
            trial_contrast[i * num_trials + j] = contrast
            np.random.shuffle(inputs)

            input_freq = np.zeros(num_groups)
            for k in range(num_groups):
                input_freq[k] = float(inputs[k]) * Hz

            file='wta.groups.%d.duration.%0.3f.p_b_e.%0.3f.p_x_e.%0.3f.p_e_e.%0.3f.p_e_i.%0.3f.p_i_i.%0.3f.p_i_e.%0.3f.contrast.%0.4f.trial.%d.h5' %\
                 (num_groups, trial_duration, wta_params.p_b_e, wta_params.p_x_e, wta_params.p_e_e, wta_params.p_e_i,
                  wta_params.p_i_i, wta_params.p_i_e, contrast, j)

            out_file = None
            if data_path is not None:
                out_file = os.path.join(data_path, file)
            wta_monitor = run_wta(wta_params,
                                  num_groups,
                                  input_freq,
                                  trial_duration,
                                  record_neuron_state=True,
                                  output_file=out_file,
                                  muscimol_amount=muscimol_amount,
                                  injection_site=injection_site,
                                  single_inh_pop=single_inh_pop)

            trial_max_bold[i * num_trials + j] = np.max(
                wta_monitor.voxel_monitor['y'].values)
            trial_max_exc_bold[i * num_trials + j] = np.max(
                wta_monitor.voxel_exc_monitor['y'].values)

    x_min = np.min(contrast_range)
    x_max = np.max(contrast_range)

    fig = plt.figure()
    clf = LinearRegression()
    clf.fit(trial_contrast, trial_max_bold)
    a = clf.coef_[0]
    b = clf.intercept_

    plt.plot(trial_contrast, trial_max_bold, 'x')
    plt.plot([x_min, x_max], [a * x_min + b, a * x_max + b], '--')
    plt.xlabel('Input Contrast')
    plt.ylabel('Max BOLD')
    plt.show()

    fig = plt.figure()
    clf = LinearRegression()
    clf.fit(trial_contrast, trial_max_exc_bold)
    a = clf.coef_[0]
    b = clf.intercept_

    plt.plot(trial_contrast, trial_max_exc_bold, 'o')
    plt.plot([x_min, x_max], [a * x_min + b, a * x_max + b], '--')
    plt.xlabel('Input Contrast')
    plt.ylabel('Max BOLD (exc only)')
    plt.show()
Example #4
0
def test_contrast_lesion(p_intra, p_inter, trial_numbers, data_path, muscimol_amount=0*nS, injection_site=0,
                         single_inh_pop=False, plot_summary=True):
    num_groups=2
    trial_duration=1.0*second

    wta_params=default_params()
    wta_params.p_b_e=0.1
    wta_params.p_x_e=0.1
    wta_params.p_e_e=p_intra
    wta_params.p_e_i=p_inter
    wta_params.p_i_i=p_intra
    wta_params.p_i_e=p_inter
    input_sum=40.0

    contrast_range=[0.0, 0.0625, 0.125, 0.25, 0.5, 1.0]
    num_trials=len(trial_numbers)
    trial_contrast=np.zeros([len(contrast_range)*num_trials,1])
    trial_max_bold=np.zeros(len(contrast_range)*num_trials)
    trial_max_exc_bold=np.zeros(len(contrast_range)*num_trials)
    for i,contrast in enumerate(contrast_range):
        print('Testing contrast %0.4f' % contrast)
        inputs=np.zeros(2)
        inputs[0]=(input_sum*(contrast+1.0)/2.0)
        inputs[1]=input_sum-inputs[0]

        for j,trial_idx in enumerate(trial_numbers):
            print('Trial %d' % trial_idx)
            trial_contrast[i*num_trials+j]=contrast
            np.random.shuffle(inputs)

            input_freq=np.zeros(num_groups)
            for k in range(num_groups):
                input_freq[k]=float(inputs[k])*Hz

            file='wta.groups.%d.duration.%0.3f.p_b_e.%0.3f.p_x_e.%0.3f.p_e_e.%0.3f.p_e_i.%0.3f.p_i_i.%0.3f.p_i_e.%0.3f.contrast.%0.4f.trial.%d.h5' %\
                 (num_groups, trial_duration, wta_params.p_b_e, wta_params.p_x_e, wta_params.p_e_e, wta_params.p_e_i,
                  wta_params.p_i_i, wta_params.p_i_e, contrast, trial_idx)

            out_file=None
            if not data_path is None:
                out_file=os.path.join(data_path,file)
            wta_monitor=run_wta(wta_params, num_groups, input_freq, trial_duration, output_file=out_file,
                single_inh_pop=single_inh_pop, record_spikes=False, record_lfp=False, save_summary_only=True)

            trial_max_bold[i*num_trials+j]=np.max(wta_monitor.voxel_monitor['y'].values)
            trial_max_exc_bold[i*num_trials+j]=np.max(wta_monitor.voxel_exc_monitor['y'].values)

    lesioned_trial_max_bold=np.zeros(len(contrast_range)*num_trials)
    lesioned_trial_max_exc_bold=np.zeros(len(contrast_range)*num_trials)
    for i,contrast in enumerate(contrast_range):
        print('Testing contrast %0.4f' % contrast)
        inputs=np.zeros(2)
        inputs[0]=(input_sum*(contrast+1.0)/2.0)
        inputs[1]=input_sum-inputs[0]

        for j,trial_idx in enumerate(trial_numbers):
            print('Trial %d' % j)
            trial_contrast[i*num_trials+j]=contrast
            np.random.shuffle(inputs)

            input_freq=np.zeros(num_groups)
            for k in range(num_groups):
                input_freq[k]=float(inputs[k])*Hz

            file='lesioned.wta.groups.%d.duration.%0.3f.p_b_e.%0.3f.p_x_e.%0.3f.p_e_e.%0.3f.p_e_i.%0.3f.p_i_i.%0.3f.p_i_e.%0.3f.contrast.%0.4f.trial.%d.h5' %\
                 (num_groups, trial_duration, wta_params.p_b_e, wta_params.p_x_e, wta_params.p_e_e, wta_params.p_e_i,
                  wta_params.p_i_i, wta_params.p_i_e, contrast, trial_idx)

            out_file=None
            if not data_path is None:
                out_file=os.path.join(data_path,file)
            wta_monitor=run_wta(wta_params, num_groups, input_freq, trial_duration, output_file=out_file,
                muscimol_amount=muscimol_amount, injection_site=injection_site, single_inh_pop=single_inh_pop,
                record_spikes=False, record_lfp=False, save_summary_only=True)

            lesioned_trial_max_bold[i*num_trials+j]=np.max(wta_monitor.voxel_monitor['y'].values)
            lesioned_trial_max_exc_bold[i*num_trials+j]=np.max(wta_monitor.voxel_exc_monitor['y'].values)

    if plot_summary:
        x_min=np.min(contrast_range)
        x_max=np.max(contrast_range)

        fig=plt.figure()
        control_clf=LinearRegression()
        control_clf.fit(trial_contrast,trial_max_bold)
        control_a=control_clf.coef_[0]
        control_b=control_clf.intercept_

        lesion_clf=LinearRegression()
        lesion_clf.fit(trial_contrast,lesioned_trial_max_bold)
        lesion_a=lesion_clf.coef_[0]
        lesion_b=lesion_clf.intercept_

        plt.plot(trial_contrast, trial_max_bold, 'xb')
        plt.plot(trial_contrast, lesioned_trial_max_bold, 'xr')
        plt.plot([x_min,x_max],[control_a*x_min+control_b,control_a*x_max+control_b],'--b',label='Control')
        plt.plot([x_min,x_max],[lesion_a*x_min+lesion_b,lesion_a*x_max+lesion_b],'--r',label='Lesioned')
        plt.xlabel('Input Contrast')
        plt.ylabel('Max BOLD')
        plt.legend()
        plt.show()

        fig=plt.figure()
        control_exc_clf=LinearRegression()
        control_exc_clf.fit(trial_contrast,trial_max_exc_bold)
        control_exc_a=control_exc_clf.coef_[0]
        control_exc_b=control_exc_clf.intercept_

        lesion_exc_clf=LinearRegression()
        lesion_exc_clf.fit(trial_contrast,lesioned_trial_max_exc_bold)
        lesion_exc_a=lesion_exc_clf.coef_[0]
        lesion_exc_b=lesion_exc_clf.intercept_

        plt.plot(trial_contrast, trial_max_exc_bold, 'ob')
        plt.plot(trial_contrast, lesioned_trial_max_exc_bold, 'or')
        plt.plot([x_min,x_max],[control_exc_a*x_min+control_exc_b,control_exc_a*x_max+control_exc_b],'--b',label='Control')
        plt.plot([x_min,x_max],[lesion_exc_a*x_min+lesion_exc_b,lesion_exc_a*x_max+lesion_exc_b],'--r',label='Lesioned')
        plt.xlabel('Input Contrast')
        plt.ylabel('Max BOLD (exc only)')
        plt.legend()
        plt.show()
Example #5
0
def test_contrast_lesion(p_intra,
                         p_inter,
                         trial_numbers,
                         data_path,
                         muscimol_amount=0 * nS,
                         injection_site=0,
                         single_inh_pop=False,
                         plot_summary=True):
    num_groups = 2
    trial_duration = 1.0 * second

    wta_params = default_params()
    wta_params.p_b_e = 0.1
    wta_params.p_x_e = 0.1
    wta_params.p_e_e = p_intra
    wta_params.p_e_i = p_inter
    wta_params.p_i_i = p_intra
    wta_params.p_i_e = p_inter
    input_sum = 40.0

    contrast_range = [0.0, 0.0625, 0.125, 0.25, 0.5, 1.0]
    num_trials = len(trial_numbers)
    trial_contrast = np.zeros([len(contrast_range) * num_trials, 1])
    trial_max_bold = np.zeros(len(contrast_range) * num_trials)
    trial_max_exc_bold = np.zeros(len(contrast_range) * num_trials)
    for i, contrast in enumerate(contrast_range):
        print('Testing contrast %0.4f' % contrast)
        inputs = np.zeros(2)
        inputs[0] = (input_sum * (contrast + 1.0) / 2.0)
        inputs[1] = input_sum - inputs[0]

        for j, trial_idx in enumerate(trial_numbers):
            print('Trial %d' % trial_idx)
            trial_contrast[i * num_trials + j] = contrast
            np.random.shuffle(inputs)

            input_freq = np.zeros(num_groups)
            for k in range(num_groups):
                input_freq[k] = float(inputs[k]) * Hz

            file='wta.groups.%d.duration.%0.3f.p_b_e.%0.3f.p_x_e.%0.3f.p_e_e.%0.3f.p_e_i.%0.3f.p_i_i.%0.3f.p_i_e.%0.3f.contrast.%0.4f.trial.%d.h5' %\
                 (num_groups, trial_duration, wta_params.p_b_e, wta_params.p_x_e, wta_params.p_e_e, wta_params.p_e_i,
                  wta_params.p_i_i, wta_params.p_i_e, contrast, trial_idx)

            out_file = None
            if not data_path is None:
                out_file = os.path.join(data_path, file)
            wta_monitor = run_wta(wta_params,
                                  num_groups,
                                  input_freq,
                                  trial_duration,
                                  output_file=out_file,
                                  single_inh_pop=single_inh_pop,
                                  record_spikes=False,
                                  record_lfp=False,
                                  save_summary_only=True)

            trial_max_bold[i * num_trials + j] = np.max(
                wta_monitor.voxel_monitor['y'].values)
            trial_max_exc_bold[i * num_trials + j] = np.max(
                wta_monitor.voxel_exc_monitor['y'].values)

    lesioned_trial_max_bold = np.zeros(len(contrast_range) * num_trials)
    lesioned_trial_max_exc_bold = np.zeros(len(contrast_range) * num_trials)
    for i, contrast in enumerate(contrast_range):
        print('Testing contrast %0.4f' % contrast)
        inputs = np.zeros(2)
        inputs[0] = (input_sum * (contrast + 1.0) / 2.0)
        inputs[1] = input_sum - inputs[0]

        for j, trial_idx in enumerate(trial_numbers):
            print('Trial %d' % j)
            trial_contrast[i * num_trials + j] = contrast
            np.random.shuffle(inputs)

            input_freq = np.zeros(num_groups)
            for k in range(num_groups):
                input_freq[k] = float(inputs[k]) * Hz

            file='lesioned.wta.groups.%d.duration.%0.3f.p_b_e.%0.3f.p_x_e.%0.3f.p_e_e.%0.3f.p_e_i.%0.3f.p_i_i.%0.3f.p_i_e.%0.3f.contrast.%0.4f.trial.%d.h5' %\
                 (num_groups, trial_duration, wta_params.p_b_e, wta_params.p_x_e, wta_params.p_e_e, wta_params.p_e_i,
                  wta_params.p_i_i, wta_params.p_i_e, contrast, trial_idx)

            out_file = None
            if not data_path is None:
                out_file = os.path.join(data_path, file)
            wta_monitor = run_wta(wta_params,
                                  num_groups,
                                  input_freq,
                                  trial_duration,
                                  output_file=out_file,
                                  muscimol_amount=muscimol_amount,
                                  injection_site=injection_site,
                                  single_inh_pop=single_inh_pop,
                                  record_spikes=False,
                                  record_lfp=False,
                                  save_summary_only=True)

            lesioned_trial_max_bold[i * num_trials + j] = np.max(
                wta_monitor.voxel_monitor['y'].values)
            lesioned_trial_max_exc_bold[i * num_trials + j] = np.max(
                wta_monitor.voxel_exc_monitor['y'].values)

    if plot_summary:
        x_min = np.min(contrast_range)
        x_max = np.max(contrast_range)

        fig = plt.figure()
        control_clf = LinearRegression()
        control_clf.fit(trial_contrast, trial_max_bold)
        control_a = control_clf.coef_[0]
        control_b = control_clf.intercept_

        lesion_clf = LinearRegression()
        lesion_clf.fit(trial_contrast, lesioned_trial_max_bold)
        lesion_a = lesion_clf.coef_[0]
        lesion_b = lesion_clf.intercept_

        plt.plot(trial_contrast, trial_max_bold, 'xb')
        plt.plot(trial_contrast, lesioned_trial_max_bold, 'xr')
        plt.plot(
            [x_min, x_max],
            [control_a * x_min + control_b, control_a * x_max + control_b],
            '--b',
            label='Control')
        plt.plot([x_min, x_max],
                 [lesion_a * x_min + lesion_b, lesion_a * x_max + lesion_b],
                 '--r',
                 label='Lesioned')
        plt.xlabel('Input Contrast')
        plt.ylabel('Max BOLD')
        plt.legend()
        plt.show()

        fig = plt.figure()
        control_exc_clf = LinearRegression()
        control_exc_clf.fit(trial_contrast, trial_max_exc_bold)
        control_exc_a = control_exc_clf.coef_[0]
        control_exc_b = control_exc_clf.intercept_

        lesion_exc_clf = LinearRegression()
        lesion_exc_clf.fit(trial_contrast, lesioned_trial_max_exc_bold)
        lesion_exc_a = lesion_exc_clf.coef_[0]
        lesion_exc_b = lesion_exc_clf.intercept_

        plt.plot(trial_contrast, trial_max_exc_bold, 'ob')
        plt.plot(trial_contrast, lesioned_trial_max_exc_bold, 'or')
        plt.plot([x_min, x_max], [
            control_exc_a * x_min + control_exc_b,
            control_exc_a * x_max + control_exc_b
        ],
                 '--b',
                 label='Control')
        plt.plot([x_min, x_max], [
            lesion_exc_a * x_min + lesion_exc_b,
            lesion_exc_a * x_max + lesion_exc_b
        ],
                 '--r',
                 label='Lesioned')
        plt.xlabel('Input Contrast')
        plt.ylabel('Max BOLD (exc only)')
        plt.legend()
        plt.show()