diff --git a/alvr/client_openxr/src/extra_extensions/multimodal_input.rs b/alvr/client_openxr/src/extra_extensions/multimodal_input.rs
index d41a5b5072..a6076be907 100644
--- a/alvr/client_openxr/src/extra_extensions/multimodal_input.rs
+++ b/alvr/client_openxr/src/extra_extensions/multimodal_input.rs
@@ -20,39 +20,37 @@ static TYPE_SIMULTANEOUS_HANDS_AND_CONTROLLERS_TRACKING_PAUSE_INFO_META: Lazy<xr
     Lazy::new(|| xr::StructureType::from_raw(1000532003));
 
 #[repr(C)]
-pub struct SystemSymultaneousHandsAndControllersPropertiesMETA {
+struct SystemSymultaneousHandsAndControllersPropertiesMETA {
     ty: xr::StructureType,
     next: *const c_void,
     supports_simultaneous_hands_and_controllers: sys::Bool32,
 }
 
 #[repr(C)]
-pub struct SimultaneousHandsAndControllersTrackingResumeInfoMETA {
+struct SimultaneousHandsAndControllersTrackingResumeInfoMETA {
     ty: xr::StructureType,
     next: *const c_void,
 }
 #[repr(C)]
-pub struct SimultaneousHandsAndControllersTrackingPauseInfoMETA {
+struct SimultaneousHandsAndControllersTrackingPauseInfoMETA {
     ty: xr::StructureType,
     next: *const c_void,
 }
 
-pub type ResumeSimultaneousHandsAndControllersTrackingMETA =
-    unsafe extern "system" fn(
-        sys::Session,
-        *const SimultaneousHandsAndControllersTrackingResumeInfoMETA,
-    ) -> sys::Result;
-pub type PauseSimultaneousHandsAndControllersTrackingMETA =
-    unsafe extern "system" fn(
-        sys::Session,
-        *const SimultaneousHandsAndControllersTrackingPauseInfoMETA,
-    ) -> sys::Result;
+type ResumeSimultaneousHandsAndControllersTrackingMETA = unsafe extern "system" fn(
+    sys::Session,
+    *const SimultaneousHandsAndControllersTrackingResumeInfoMETA,
+) -> sys::Result;
+type PauseSimultaneousHandsAndControllersTrackingMETA = unsafe extern "system" fn(
+    sys::Session,
+    *const SimultaneousHandsAndControllersTrackingPauseInfoMETA,
+) -> sys::Result;
 
 pub struct MultimodalMeta {
-    pub session: xr::Session<xr::AnyGraphics>,
-    pub resume_simultaneous_hands_and_controllers_tracking_meta:
+    session: xr::Session<xr::AnyGraphics>,
+    resume_simultaneous_hands_and_controllers_tracking_meta:
         ResumeSimultaneousHandsAndControllersTrackingMETA,
-    pub pause_simultaneous_hands_and_controllers_tracking_meta:
+    pause_simultaneous_hands_and_controllers_tracking_meta:
         PauseSimultaneousHandsAndControllersTrackingMETA,
 }
 
diff --git a/alvr/client_openxr/src/interaction.rs b/alvr/client_openxr/src/interaction.rs
index b54e1e1a4a..e9c2eaa17e 100644
--- a/alvr/client_openxr/src/interaction.rs
+++ b/alvr/client_openxr/src/interaction.rs
@@ -108,7 +108,8 @@ pub struct InteractionContext {
     pub action_set: xr::ActionSet,
     pub button_actions: HashMap<u64, ButtonAction>,
     pub hands_interaction: [HandInteraction; 2],
-    pub multimodal_hands: Option<(MultimodalMeta, bool)>,
+    multimodal_handle: Option<MultimodalMeta>,
+    pub multimodal_hands_enabled: bool,
     pub face_sources: FaceSources,
     pub body_sources: BodySources,
 }
@@ -229,13 +230,13 @@ impl InteractionContext {
             "/user/hand/right/output/haptic",
         ));
 
-        let multimodal = create_ext_object("MultimodalMeta", Some(true), || {
-            MultimodalMeta::new(xr_session.clone(), &extra_extensions, xr_system)
+        let multimodal_handle = create_ext_object("MultimodalMeta", Some(true), || {
+            MultimodalMeta::new(xr_session.clone(), extra_extensions, xr_system)
         });
 
         let left_detached_controller_pose_action;
         let right_detached_controller_pose_action;
-        if multimodal.is_some() {
+        if multimodal_handle.is_some() {
             // Note: when multimodal input is enabled, both controllers and hands will always be active.
             // To be able to detect when controllers are actually held, we have to register detached
             // controllers pose; the controller pose will be diverted to the detached controllers when
@@ -359,7 +360,8 @@ impl InteractionContext {
                     skeleton_tracker: right_hand_tracker,
                 },
             ],
-            multimodal_hands: multimodal.map(|m| (m, false)),
+            multimodal_handle,
+            multimodal_hands_enabled: false,
             face_sources: FaceSources {
                 combined_eyes_source,
                 eye_tracker_fb: None,
@@ -375,12 +377,10 @@ impl InteractionContext {
 
     pub fn select_sources(&mut self, config: &InteractionSourcesConfig) {
         // First of all, disable/delete all sources. This ensures there are no conflicts
-        if let Some((handle, enabled)) = &mut self.multimodal_hands {
-            if *enabled {
-                handle.pause().ok();
-                *enabled = false;
-            }
+        if let Some(handle) = &mut self.multimodal_handle {
+            handle.pause().ok();
         }
+        self.multimodal_hands_enabled = false;
         self.face_sources.eye_tracker_fb = None;
         self.face_sources.face_tracker_fb = None;
         self.face_sources.eye_tracker_htc = None;
@@ -415,10 +415,9 @@ impl InteractionContext {
         // Note: We cannot enable multimodal if fb body tracking is active. It would result in a
         // ERROR_RUNTIME_FAILURE crash.
         if config.body_tracking.is_none() && config.prefers_multimodal_input {
-            if let Some((handle, enabled)) = &mut self.multimodal_hands {
-                if !*enabled {
-                    handle.resume().ok();
-                    *enabled = true;
+            if let Some(handle) = &mut self.multimodal_handle {
+                if handle.resume().is_ok() {
+                    self.multimodal_hands_enabled = true;
                 }
             }
         }
diff --git a/alvr/client_openxr/src/stream.rs b/alvr/client_openxr/src/stream.rs
index d029549806..e789b886ec 100644
--- a/alvr/client_openxr/src/stream.rs
+++ b/alvr/client_openxr/src/stream.rs
@@ -510,17 +510,12 @@ fn stream_input_loop(
 
         // Note: When multimodal input is enabled, we are sure that when free hands are used
         // (not holding controllers) the controller data is None.
-        let multimodal_enabled = int_ctx
-            .multimodal_hands
-            .as_ref()
-            .map(|(_, enabled)| *enabled)
-            .unwrap_or(false);
-        if multimodal_enabled || left_hand_skeleton.is_none() {
+        if int_ctx.multimodal_hands_enabled || left_hand_skeleton.is_none() {
             if let Some(motion) = left_hand_motion {
                 device_motions.push((*HAND_LEFT_ID, motion));
             }
         }
-        if multimodal_enabled || right_hand_skeleton.is_none() {
+        if int_ctx.multimodal_hands_enabled || right_hand_skeleton.is_none() {
             if let Some(motion) = right_hand_motion {
                 device_motions.push((*HAND_RIGHT_ID, motion));
             }