Virtual Reality simulator for dental anesthesia training in the inferior alveolar nerve block

Abstract Objectives This study shows the development and validation of a dental anesthesia-training simulator, specifically for the inferior alveolar nerve block (IANB). The system developed provides the tactile sensation of inserting a real needle in a human patient, using Virtual Reality (VR) techniques and a haptic device that can provide a perceived force feedback in the needle insertion task during the anesthesia procedure. Material and Methods To simulate a realistic anesthesia procedure, a Carpule syringe was coupled to a haptic device. The Volere method was used to elicit requirements from users in the Dentistry area; Repeated Measures Two-Way ANOVA (Analysis of Variance), Tukey post-hoc test and averages for the results’ analysis. A questionnaire-based subjective evaluation method was applied to collect information about the simulator, and 26 people participated in the experiments (12 beginners, 12 at intermediate level, and 2 experts). The questionnaire included profile, preferences (number of viewpoints, texture of the objects, and haptic device handler), as well as visual (appearance, scale, and position of objects) and haptic aspects (motion space, tactile sensation, and motion reproduction). Results The visual aspect was considered appropriate and the haptic feedback must be improved, which the users can do by calibrating the virtual tissues’ resistance. The evaluation of visual aspects was influenced by the participants’ experience, according to ANOVA test (F=15.6, p=0.0002, with p<0.01). The user preferences were the simulator with two viewpoints, objects with texture based on images and the device with a syringe coupled to it. Conclusion The simulation was considered thoroughly satisfactory for the anesthesia training, considering the needle insertion task, which includes the correct insertion point and depth, as well as the perception of tissues resistances during the insertion.


Introduction
In light of the importance of training in the health eld, including practical classes to teach medical and clinical procedures, the use of appropriate tools is crucial. Such tools can be mannequin models and other tangible objects, or computer systems that allow virtual simulations, called simulators. In addition, there are hybrid systems that combine different elements (real and virtual) to offer to students experiences that are close to those found in the clinical practice.
Simulation-based medical training tools developed via Virtual Reality (VR) techniques have increased in recent years 16 . These techniques provide threedimensional virtual environments with real-time interaction, which include different degrees of immersion and realism, thereby allowing an accurate reproduction of tasks that are essential for health care training procedures.
VR simulating training systems, especially those involving the manipulation of medical instruments, generally use haptic interfaces. The term haptic originates from the Greek word haptesthai, which means "to touch" 25 . The tactile sense consists of sensations triggered when the skin is subjected to mechanical, thermal, or chemical stimuli 4 . The most common kind of haptic interface used provides force feedback, i.e, this kind allows simulating procedures that require perfect force control or pressure executed by the student.
Within the context described, this study presents the development and validation of a haptic-based VR anesthesia injection training simulator for Dentistry procedures, speci cally the local anesthesia, for the inferior alveolar nerve block (IANB), the most common type of nerve block used for dental procedures.
Furthermore, IANB has a high failure rate, approximately 20-25% 13 . These failures can be classi ed as: anatomical (changes in nerve pathways and mandibular foramen), pathological (infection, previous surgeries, and in ammation), pharmacological (alcoholism and the use of certain substances), physiological (fear and anxiety) and also inadequate technique (improper needle insertion and beginning treatment too quickly) 13,19 .
The literature has several studies published on needle insertion simulation, such as: prostate brachytherapy 9 , cell injection 17 , types of biopsies 8,20 and other types of anesthesia 11 . However, thus far there is only one simulator to train the administration of dental anesthesia, similar to the one developed in this study, which emphasizes visualization compared with haptic interaction 23 . The objective of this article is to present the development and validation of a dental anesthesia training simulator, speci cally for the IANB.
The system considers the anatomical characteristics of children aged 7 to 12 years old and the direct anesthetic technique, focusing on the inferior alveolar nerve. The simulator provides the tactile sensation of inserting a real needle in a human patient, using VR techniques and a haptic device that provides a force feedback in the needle insertion task during the anesthesia procedure.

Material and methods
The simulator was developed considering the following steps: (1) requirements analysis for de ning the computational characteristics and literature review; (2) simulator development; (3) evaluation of the simulator; (4) analysis of the results.
The Volere method 24 was used to gather the system requirements from interviews with teachers and students from a dental education institution. A systematic review method was applied for the literature review 15 , for data collection of the state of the art regarding training tools based on VR simulation with haptic feedback.
The haptic device used (Phantom Omni * ) captures the movements performed by the student, including rotation and position, which are then reproduced in a three-dimensional virtual environment. The device also captures the force applied to the simulator by the student and returns tactile sensations, which favor the perceived force of the virtual components involved in the training.
A subjective evaluation was applied, using questionnaires to collect information about the tool in the opinion of dentists and future dentists. The questionnaire was detailed in subsection Questionnaires and included pro le, preferences (number of viewpoints, texture of the objects, and haptic device handler), as well as visual (appearance, scale, and position of objects), and haptic aspects (motion space, tactile sensation, and motion reproduction).

Requirements
The requirements for the training simulator were: (1) display three-dimensional objects to realistically represent the anatomical structures (teeth, jaw, gums, tongue, facial skin) and dental instruments,

Three-dimensional modeling
To simulate the virtual patient and the components involved in the training procedure, the following objects were modeled: facial skin, teeth, tongue, muscles, nerves, blood vessels, and bones.
The objects were devised to ensure realism of the actual structures, that is, considering attributes such as size, shape and textures of the anatomical structures, The structures must be similar to real patients (in this case, children of 7 and 12 years of age).
The colors for virtual structures are de ned using illumination parameters available in the simulator, as well as size and position of the objects. The images for texture were created capturing real images (to represent bones, for example) and images from anatomy books. The skin image was detailed with pores and oiliness. These books were also used to de ne the shape of the structures.
Objects, such as the skin, can be added and removed using the keyboard. This functionality is available for students who desire to view the internal structures, which, in the actual procedure, is partially obtained by retracting the cheek.

Collision detection
The collision detection adds a complexity to the simulation, because detailed three-dimensional objects

Objective and hypotheses
The main objective of the experiments was validating the VR simulator for dental anesthesia training, the IANB procedure, especially the needle insertion task.
The following hypotheses were established: and at the correct depth (needle tip next to the nerve) to IANB on the right side, without reaching the nerve.
After performing the task in all the scenarios, the participant answered the questionnaire.

Questionnaires
Questionnaires were designed to be applied after the tests, to collect pro le, preferences, visual, and haptic aspects. They also were used to verify the participant's sensations, such as exhaustion and understanding of the task to be performed in the virtual environment. The participants could offer suggestions, assessment, and ideas to improve the simulator. The preferences consisted of information on the  Repeated Measures Two-Way ANOVA was applied separately for each aspect (Visual and Haptic). We used the following conventional assumptions for an ANOVA test: the samples must be random and independent, the data must have a normal distribution, and the groups' variances must be equal. Since the  A p-value less than 0.01, or value less than 1%, indicates which value is signi cant at the 99% level.
Thus, the probability of average difference between the responses of users to occur by chance is less than 1%.   (cadavers and animals) in training situations involves ethical issues that must be properly considered, as well as the dif culties to obtain these materials for training 3,5 . In the virtual training context, realism is very important 5,9,20,23  As for hypothesis 2 (texture of three-dimensional objects), the preferred objects were those based on images, seen as more realistic, however, some participants appreciated the objects with colors. As for the texture of objects, the preferred one was based on image, and all the experts considered objects based on texture as the most suitable choice, however, beginners and intermediates gave relatively high opinion scores to color.
Regarding hypothesis 3 (the haptic device handler), manipulation via Carpule syringe was the one preferred because the syringe is the instrument routinely used by dentists to administrate anesthesia.
As for the device (original or adapted), the adapted device was the preferred one, receiving high values in all the groups of levels of experience. Some beginners, with no experience, preferred the original stylus, which can be explained by the lack of experience with the Carpule syringe. It was observed that how the device is held during the needle manipulation depends on the object format (stylus or syringe), as shown in Figure 1, and holding the device handler like a pen to perform the anesthesia procedure is wrong.
Regarding hypothesis 4 (scale, position, and object format) and hypothesis 5 (space to move the device handler, motion reproduction and force feedback), which include the visual and haptic aspects, respectively, the participants preferred the virtual objects, with scale, position and formats suitable for training. They reported that the haptic feedback should be improved, with speci c corrections, especially in the tactile feedback of anatomical structures, such as the mucosa. The workspace of the device, and also motion reproduction were considered satisfactory.
However, other trajectories and motions will be studied, as well as other ways for improving the human-computer interaction with the haptic device used in the experiments for dental anesthesia training.
However, the VR simulator allows the user to calibrate the force feedback, specifying resistance values for each type of anatomical structure: skin, mucosa, tongue, gums, bones, muscles, and nerves.
The haptic perception data of the research will be used to re ne the resistance values of the structures, which will be evaluated in future tests. The maximum force feedback value provided by the device was not considered as being insuf cient. Although the device does not have six degrees of freedom of force feedback, the participants did not report problems.
A question which deserves to be highlighted is that the simulator presented here considers the anatomy of children. When patients are children, the foramen is usually located slightly below the occlusal plane 19 , which is different in adult patients. Thus, the procedure in children is different due to the size of the structures in these patients. However, all the methods developed can work with objects modeled to represent any type of patient, considering any age.
This rst version of the simulator can already be used for IANB training. However, the project intends to improve the system. Thus, future works include the addition of stereoscopy techniques for threedimensional imaging with volumetric perception, other anatomical structures (adults and pathologies, for example), implementation of the anesthetic injection, and the removal of the cheek using the hand, as well as adapting the simulator to transform the system into a game, using scores during the procedure to specify success and errors. CORRÊA CG, MACHADO MAAM, RANZINI E, TORI R, NUNES FLS