diff --git a/Tests/test_sklearn.py b/Tests/test_sklearn.py
index 62c6707..bd51aff 100644
--- a/Tests/test_sklearn.py
+++ b/Tests/test_sklearn.py
@@ -5,9 +5,8 @@
 # Test case using the fixture
 @pytest.mark.timeout(300)  # Extended timeout to 5 minutes
 def test_run_bci():
-    bci = PyBCI(minimumEpochsRequired=5, createPseudoDevice=True)
     clf = MLPClassifier(max_iter = 1000, solver ="lbfgs")#solver=clf, alpha=alpha,hidden_layer_sizes=hid)
-    bci = PyBCI(minimumEpochsRequired=5, createPseudoDevice=True)
+    bci = PyBCI(minimumEpochsRequired=5, createPseudoDevice=True,  clf = clf)
     while not bci.connected:
         bci.Connect()
         time.sleep(1)
diff --git a/pybci/Utils/PseudoDevice.py b/pybci/Utils/PseudoDevice.py
index f684431..07733b6 100644
--- a/pybci/Utils/PseudoDevice.py
+++ b/pybci/Utils/PseudoDevice.py
@@ -26,7 +26,7 @@ def __init__(self, execution_mode='process', *args, **kwargs):
             # Note: Don't initialize self.device here for 'process' mode!
         elif self.execution_mode == 'thread':
             self.command_queue = None  # Not needed for threads, but kept for consistency
-            self.stop_signal = multiprocessing.Event()
+            self.stop_signal = False
             self.device = PseudoDevice(*self.args, **self.kwargs, stop_signal=self.stop_signal, is_multiprocessing=False)  # Initialize for 'thread' mode
             self.worker = threading.Thread(target=self._run_device)
         else:
@@ -42,10 +42,7 @@ def __init__(self, execution_mode='process', *args, **kwargs):
             self.log_reader_process.start()
     
     def __del__(self):
-        if not self._should_stop():
-            self.stop_signal.set()
-        self.worker.join()
-
+        self.StopStreaming()  # Your existing method to stop threads and processes
 
     def _run_device(self):
         if self.execution_mode == 'process':
@@ -84,7 +81,8 @@ def StopStreaming(self):
 
         self.worker.join()  # Wait for the worker to finish
 
-        if self.log_reader_process:
+        # Terminate the log reader process if it exists
+        if self.log_reader_process is not None:
             self.log_reader_process.terminate()
             self.log_reader_process.join()
 
@@ -198,6 +196,7 @@ def update(self):
                 self.outlet.push_chunk(num.tolist())
                 self.last_update_time = current_time
 
+    # Make sure this method is available in the PseudoDevice class
     def StopStreaming(self):
         if self.is_multiprocessing:
             self.stop_signal.set()
@@ -207,10 +206,6 @@ def StopStreaming(self):
         if hasattr(self, 'thread'):
             self.thread.join()  # Wait for the thread to finish
 
-        if self.pseudoMarkerConfig.autoplay and hasattr(self, 'marker_thread'):
-            self.marker_thread.join()  # Wait for the marker thread to finish
-
-        self.log_message(Logger.INFO, " PseudoDevice - Stopped streaming.")
     def BeginStreaming(self):
         if self.is_multiprocessing:
             # For multiprocessing, we assume the worker process is already running
diff --git a/t.py b/t.py
new file mode 100644
index 0000000..5f431f4
--- /dev/null
+++ b/t.py
@@ -0,0 +1,34 @@
+from pybci import PyBCI
+import time
+
+from sklearn.neural_network import MLPClassifier
+# Test case using the fixture
+
+def test_run_bci():
+    clf = MLPClassifier(max_iter = 1000, solver ="lbfgs")#solver=clf, alpha=alpha,hidden_layer_sizes=hid)
+    bci = PyBCI(minimumEpochsRequired=5, createPseudoDevice=True)
+    while not bci.connected:
+        bci.Connect()
+        time.sleep(1)
+    bci.TrainMode()
+    accuracy_achieved = False
+    marker_received = False
+    accuracy=0
+    while True:
+        currentMarkers = bci.ReceivedMarkerCount() # check to see how many received epochs, if markers sent to close together will be ignored till done processing
+        time.sleep(0.5) # wait for marker updates
+        print("Markers received: " + str(currentMarkers) +" Accuracy: " + str(round(accuracy,2)), end="         \r")
+        if len(currentMarkers) > 1:  # check there is more then one marker type received
+            marker_received = True
+            if min([currentMarkers[key][1] for key in currentMarkers]) > bci.minimumEpochsRequired:
+                classInfo = bci.CurrentClassifierInfo() # hangs if called too early
+                accuracy = classInfo["accuracy"]###
+                if accuracy > 0.75:
+                    accuracy_achieved = True
+                    bci.StopThreads()
+                    break
+            if min([currentMarkers[key][1] for key in currentMarkers]) > bci.minimumEpochsRequired+4:
+                break
+    #assert accuracy_achieved and marker_received
+if __name__ == '__main__':
+    test_run_bci()
\ No newline at end of file