Averageover tapers

frites/conn/conn_spec.py

@@ -242,7 +242,8 @@ def conn_spec(
     kernel = _create_kernel(sm_times, sm_freqs, kernel=sm_kernel)
 
     # average over tapers
-    tapers_average = mode == 'multitaper'
+    tapers_average = False
+    # tapers_average = mode == 'multitaper'
 
     # define arguments for parallel computing
     mesg = f'Estimating pairwise {f_name} for trials %s'
@@ -339,13 +340,14 @@ def conn_spec(
                      coords=(trials, roi, times))
     freqs = np.linspace(2, 60, 40)
     n_cycles = freqs / 2.
+    mt_bandwidth = np.linspace(4, 10, len(freqs))
 
     foi = np.array([[2, 4], [5, 7], [8, 13], [13, 30], [30, 60]])
     coh = conn_spec(
         x, sfreq=sfreq, roi='roi', times='times', sm_times=2.,
         sm_freqs=1, mode='multitaper', n_cycles=n_cycles, freqs=freqs,
         decim=1, foi=None, n_jobs=1, metric='coh', mean_trials=False,
-        block_size=2, **kw_links
+        block_size=2, mt_bandwidth=None, **kw_links
     )
 
     # coh.mean(('trials', 'roi')).plot()

frites/conn/conn_tf.py

@@ -66,24 +66,23 @@ def _tf_decomp(data, sf, freqs, mode='morlet', n_cycles=7.0, mt_bandwidth=None,
         # the MT decomposition is done separatedly for each
         # Frequency center
         if isinstance(mt_bandwidth, (list, tuple, np.ndarray)):
-            raise NotImplementedError("Not compatible with multiple bandwidth")
-            # # Arrays freqs, n_cycles, mt_bandwidth should have the same size
-            # assert len(freqs) == len(n_cycles) == len(mt_bandwidth)
-            # out = []
-            # for f_c, n_c, mt in zip(freqs, n_cycles, mt_bandwidth):
-            #     _out = tfr_array_multitaper(
-            #         data, sf, [f_c], n_cycles=float(n_c), time_bandwidth=mt,
-            #         output='complex', decim=decim, n_jobs=n_jobs, **kw_mt
-            #     )
-            #     out.append(_out)
-
-            # # stack everything
-            # out = np.concatenate(out, axis=2)
+            # Arrays freqs, n_cycles, mt_bandwidth should have the same size
+            assert len(freqs) == len(n_cycles) == len(mt_bandwidth)
+            out = []
+            for f_c, n_c, mt in zip(freqs, n_cycles, mt_bandwidth):
+                _out = tfr_array_multitaper(
+                    data, sf, [f_c], n_cycles=float(n_c), time_bandwidth=mt,
+                    output='complex', decim=decim, n_jobs=n_jobs, **kw_mt
+                )
+                out.append(_out.mean(2))
+            out = np.concatenate(out, axis=2)
         elif isinstance(mt_bandwidth, (type(None), int, float)):
             out = tfr_array_multitaper(
                 data, sf, freqs, n_cycles=n_cycles,
                 time_bandwidth=mt_bandwidth, output='complex', decim=decim,
                 n_jobs=n_jobs, **kw_mt)
+            # mean across tapers
+            out = out.mean(axis=2)
     else:
         raise ValueError('Method should be either "morlet" or "multitaper"')