The BIM Neural Pathways (BNP) concept: BIM competence pathways reinforced by micro-loops in real project events

The BIM Neural Pathways (BNP) concept: BIM competence pathways reinforced by micro-loops in real project events Andrzej Szymon Borkowski, DSc PhD Eng. Department of Spatial Planning and Environmental Sciences, Faculty of Geodesy and Cartography, Warsaw University of Technology, Warsaw, Poland ORCID: 0000-0002-7013-670X email: [email protected] ABSTRACT: The paper presents the concept of BIM Neural Pathways (BNP) as a complementary approach to the existing BIM Body of Knowledge framework for defining and developing BIM competencies. Instead of focusing on static lists of use cases and KSAs, BNP places competencies within the course of specific project events (e.g., BIM kickoff, model intake, coordination sprints, stage freeze, 4D/5D baseline, FM handover), treating them as dynamic activity patterns reinforced by repetitive organisational stimuli and project data. The concept of a ‘BIM competency neuron’ is proposed as the smallest unit of competency visible in the behaviour of team members and in data evidence (CDE issues, QA/QC reports, decision logs). Each neuron is described in a standardised card, including business intent, behavioural indicators, required data evidence, proficiency levels, prerequisites, and synaptic connections to other neurons. The neurons are used to build a competency graph and an ‘event-neuron’ matrix, allowing the identification of competency risk areas, central neurons, and configurations specific to project roles. The development mechanism is based on a micro-loop embedded directly in the course of an event (prebrief, action, capture, micro-drill, level update). It was indicated how micro-loops can be integrated with existing BIM processes and how data from CDE and project dashboards can be used to measure neuron activation and their impact on project KPIs (time, cost, rework, coordination quality). BNP does not replace classic competency frameworks, but provides the missing mechanism of continuous, micro learning ‘on the job’, which is particularly important in the context of the rapid development of BIM and AI technologies. KEYWORDS: BIM Neural Pathways (BNP), BIM competencies, BIM Body of Knowledge (BIM BOK), microlearning, micro-loop, project events 1. INTRODUCTION In recent years, BIM Body of Knowledge (BIM BOK) has become one of the most important reference points for how the industry describes and develops competencies in BIM (Building Information Modeling). Its creators, working within the Academic Interoperability Coalition (AiC), proposed a consistent language of competencies based on BIM use cases, defined professional tasks, and KSA (Knowledge, Skills, Abilities) sets broken down into proficiency levels from entry level to full independence (Wu et al., 2018a; Wu et al., 2018b). Thanks to this, BIM BOK has made it possible to compare curricula between universities, standardize job profiles in companies, and design training and certification based on common competency criteria (Wu et al., 2018a). An important development of this concept is the book Developing BIM Talent, which shows how to transform BIM BOK into practical education and staff development programs, with assessment metrics and clearly assigned KSAs to roles and career stages (Wu et al., 2021). In practice, this is a milestone because, for the first time, BIM has been described not in terms of specific tools, but in terms of what people actually do in the investment process and what competencies are needed to do so (Wu et al., 2018b). At the same time, BIM BOK grows out of the logic of a certain knowledge map, which catalogs competencies in the form of a relatively stable list of use cases and KSAs and assumes that talent development consists of successively "covering" this list in education or company training (Wu et al., 2021). This approach works well when an organization needs to organize competency requirements, prepare formal role profiles, or implement BIM from scratch. However, as BIM matures, the limitations of this concept also grow. BIM roles and applications are evolving faster than standards, and the real differences in quality between teams are increasingly not about whether they know the use case, but how they behave at critical moments in the project: under time pressure, in the face of rapid design changes, or in situations of conflict of interest between industries. BIM BOK therefore describes competencies mainly as acquired (a list of KSA to be mastered), and less as consolidated or reinforced in action in the rhythm of the investment and construction project's . The new BIM Neural Pathways (BNP) concept presented in this article is being developed as a conscious competitor to this approach. It does not negate BIM BOK; on the contrary, it starts from the identification of the same problems: a shortage of BIM personnel, scattered competency standards, and a rapidly growing number of BIM applications in AECOO (Architecture, Engineering Construction Owners Operators) (Wu et al., 2018a; Wu et al., 2021). However, it proposes a different methodology. Where BIM BOK organizes the process through a list of use cases, BNP organizes it through project events, i.e., repeatable, real-life events from the life of an investment, such as BIM kickoff, model intake, coordination sprints, stage freeze, 4D/5D baseline, as-built handover, or asset data transfer to FM. It is these events that are the stimuli that activate competencies in practice: they force decisions, expose gaps, test cooperation, and leave a trace in the project data. In this sense, BNP shifts the focus from the question "what KSAs do you need to know?" to the question "what competencies need to be activated in a specific event so that the project does not enter a spiral of rework, delays, or disputes?" The second axis of competitiveness is the development mechanism. BIM BOK supports modular education: courses, thematic blocks, and certifications based on mapping KSA to use cases (Wu et al., 2021). BIM Neural Pathways, on the other hand, assumes that competencies are like neural pathways: they are created and reinforced by repetitive impulses in a real work environment. Therefore, the basic unit of learning is not a course, but a socalled "micro-loop" of learning embedded in a project event. A micro-loop is a short cycle consisting of: • pre-brief (goal and expected standard), • action in the project, • capture (what happened), • micro-drill (brief correction of knowledge through accommodation or practice through assimilation), • proficiency level update. Each such cycle reinforces the competency path exactly where it is needed, i.e., in the rhythm of real work, rather than in a training room or lecture hall detached from the context. This fundamentally changes the logic of education, as development ceases to be a process external to the project and becomes a mechanism built into its regular events. The third difference concerns what is measured and how proficiency is verified. BIM BOK proposes competency levels and metrics based on KSA and performance levels (Wu et al., 2021). BNP goes further in the direction of evidence from practice, i.e., proficiency levels that are updated based on what actually remains after an event, such as issue history (BCF/RFI), decision logs in CDE, QA/QC reports, or recurring problem indicators. Competence ceases to be a declaration or test result and becomes an observable, auditable trace in project data. This makes it possible to build a team competency matrix, a reliable profile of neural pathways associated with project risks and future events, which supports real-time personnel planning and mentoring. BIM Neural Pathways (BNP) does not replace BIM BOK, but competes with it at the level of talent development philosophy. BIM BOK is a map of knowledge and tasks that tells you what you need to know and greatly stabilizes the language of the industry (Wu et al., 2018a). BNP is a map of competency plasticity that tells you how people actually learn BIM in a project, how to strengthen those competencies with the micro-loops discussed, and how to scale that process in the face of constant changes in technology and AECOO practice. If BIM BOK is a solid foundation for the standardization of competencies, then BIM Neural Pathways is a system for their continuous generation and consolidation (dynamic, event-based, practice-based, and evidence-based). In the era of dynamic development of AI (Artificial Intelligence), this is important because the development of so-called critical technologies is so dynamic that it is difficult to keep up with changing trends. Thus, traditional courses and training are not enough. Employees must learn on an ongoing basis during their daily professional practice. 2. NEURONS AND NEURAL PATHWAYS BIM COMPETENCES The purpose of this chapter is to present the methodological core of the BIM Neural Pathways concept. The first chapter justified the need for an alternative approach to BIM staff development in relation to the BIM BOK approach, while the second chapter moves on to the operational level: it specifies the unit of competence analysis (neural pathway), shows how it is formally described, and explains the rules for building a network of dependencies and development paths. In this sense, the chapter serves a similar function to the methodological part of BIM BOK, in which the authors of define professional tasks, KSAs, and proficiency levels and validate them by expert consensus (Wu et al., 2018a; Wu et al., 2018b). The difference, however, is that BNP does not adopt KSA as the basic atom (unit) of competence, but proposes a unit that is even more granular and directly observable in the course of projects. In BNP terms, a BIM competency "neuron" is defined as the smallest unit of skill that manifests itself in recognizable behavior or in a produced project artifact. This is a direct reference to the tradition of research on individual BIM competencies, which emphasizes their measurability and the possibility of aggregating them into larger structures of organizational capabilities (Succar, Sher, & Williams, 2013). Thus, a neuron is not synonymous with proficiency in a tool (e.g., "ability to use Autodesk Revit"), a broad activity label (e.g., "competence in model coordination"), or a job title (e.g., BIM Modeler). In design practice, such aggregates often mask actual development gaps because the team may "perform coordination," but the recurrence of conflicts and rework indicates a lack of specific micro-skills, such as collision triage, proper decision ownership, or recording justifications in a CDE (Common Data Environment). BIM BOK deliberately describes competencies at the level of use cases and KSAs, as this has standardization value for education, recruitment, and certification (Wu et al., 2021). BIM Neural Pathways assumes that real-time development control requires an even more atomic level, i.e., a neuron, and a flexible level, i.e., a neural pathway that can be consolidated (myelinated like an axon) or made more flexible (like a neural connection). In order for a neuron to become a diagnostic and teaching tool, rather than just a metaphor, it must be precisely operationalized. In BNP, this is achieved by a standardized "neuron card," a recording format that enables consistent identification of competencies across different organizations and projects. The logic of the card refers to the BIM BOK formalism, where each professional task is linked to a set of KSA and performance levels (Wu et al., 2018a). However, the neuron card shifts the emphasis from "what someone knows/can do" to "how they reveal it in action" and "in which project data there is a trace of this competence." For this reason, the description of a neuron includes: (1) an operational working name; (2) the intention of the neuron, i.e., a clear definition of its value for the project; (3) a set of behavioral indicators, i.e., observable activities that allow the use of the competency to be recognized; (4) types of evidence artifacts that can be integrated with the CDE; (5) proficiency levels 0–4 described qualitatively; (6) prerequisite dependencies and (7) co-activation synapses with other neurons; (8) anti-patterns distinguishing apparent performance from competent performance; and (9) short training procedures (micro-drills) to be used in micro-learning loops. This format corresponds to the postulate of BIM competency literature that competency units should be both measurable and contextual, i.e., embedded in real information production processes (Succar et al., 2013). To illustrate the description of a neuron, we can refer to the technical and process competency "Clash Triage & Resolution" (Fig. 1). The intention of the neuron here is to reduce rework and coordination costs by prioritizing conflicts and ensuring owner decisions and an auditable trail. Behavior indicators include, among others, classifying collisions according to their impact on safety, cost, schedule, and operation, assigning an "owner of resolution," closing issues in BCF/CDE with justification, and identifying systemic causes of conflicts. Evidence artifacts include issue history (e.g., BCF), decision logs in CDE, and conflict recurrence/reopening indicators. Proficiency levels are distinguished not by the number of clashes detected, but by the quality of triage and the ability to stabilize the coordination process. The prerequisites are neurons associated with model federation and detection rule creation, and synapses connect these neurons to the facilitation of coordination sprints and crossindustry negotiations. This example shows the difference between the KSA description and the neuron, where the neuron is not "knowledge about coordination" but a measurable pattern of behavior in a specific project event. Figure1 : An example of a BIM competency neuron and its synaptic connections in the BIM Coordinator's neural pathway during a collision coordination sprint. After defining individual neurons, the concept moves to the structural level. BNP assumes that competencies in BIM function in the form of a dependency graph, analogous to how BIM BOK links KSA with use cases and professional tasks (Wu et al., 2018b; Wu et al., 2021). The difference lies in the typology of relationships. Synapses in BNP are of a prerequisite nature (neuron A determines the appearance of neuron B), co-activating (neurons cooccur in the same event), or stabilizing (social neurons stabilize the operation of technical neurons, reducing organizational friction). This typology corresponds to the difference observed in the literature between technical and informational competencies and collaboration and change management competencies, which in practice determine the BIM maturity of an organization. As a result, the neuron graph not only describes "what is needed," but also allows modeling "what reinforces what" in real events. Neuron pathways are then derived from the neuron graph, understood as the minimum and sufficient configurations of neurons that allow critical project events to be carried out with the expected quality. At this point, Neural Pathways makes a conceptual shift from BIM BOK. BIM BOK defines roles by assigning professional tasks and KSAs at performance levels (Wu et al., 2021). Neural Pathways defines roles as network phenotypes, i.e., sets of neurons connected by synapses that together enable stable execution of events. For example, the BIM Coordinator role trail includes technical neurons (model federation, collision rule design, triage, and issue creation), process neurons (CDE governance, DoD criteria definition, sprint facilitation), and social neurons (cross-industry translation, negotiations, conflict de-escalation). In such a structure, the role is not the sum of competencies, but the configuration of their interaction. The final element of the chapter is the validation of neurons and networks. BIM Neural Pathways assumes that a neuron is correctly defined only when it meets empirical criteria, i.e., it is possible to identify a project event in which the neuron is critical and leaves a recognizable evidence artifact in the project data, and its absence generates a measurable decrease in quality, time, or information security because its description does not duplicate other neurons. Such criteria are in line with both the logic of BIM competence validation by experts (Wu et al., 2018b) and the postulates of the literature on the need to closely link competences to the process of information production in the life cycle of an object (Succar et al., 2013). In practice, validation takes place iteratively in pilot projects, where it is observed whether the neuron is actually distinguishable at levels 0–4 and whether micro-drills translate into improved event flow. The second chapter thus establishes the ontology of the BIM Neural Pathways system (the neuron as a unit of competence, the synapse as a relationship of interaction, neural pathways, and the pathway as a structure of role and development). This makes it possible to design the first neuron cards, build a competence graph, and derive pathways for selected roles from it. In the following chapters, these tools have been embedded in a catalog of design events and learning micro-loops, which allows us to move from describing competencies to controlling their development in real time. 3. BIM NEURAL PATHWAYS: PROJECT EVENTS AND MICRO-LOOPS AS A MECHANISM FOR COMPETENCE DEVELOPMENT 3.1 Project events as stimuli for BIM competency pathways In the BIM Neural Pathways concept, a project event is the basic unit of situational analysis in which BIM competency neurons are activated, strengthened, or disintegrated. An event is understood here as a repeatable, easily recognizable event in the life cycle of a construction project, such as BIM kickoff, model intake, coordination sprint, stage freeze, 4D/5D baseline, as-built handover, or transfer of asset data to FM (Fig. 2). It is at these critical points in the investment process that information requirements, time pressure, conflicts of interest, and technological limitations converge, i.e., precisely those factors that reveal the actual state of the BIM team's competence. From the perspective of traditional BIM BOK, project events remained in the background, as the main organizing units were use cases and the corresponding professional tasks and KSAs, broken down into proficiency levels (Wu et al., 2018a; Wu et al., 2018b; Wu et al., 2021). Meanwhile, the practice of mature BIM organizations shows that the success of implementation is determined not only by whether the team has "mastered" a given use case, but also by how it behaves at critical moments of the project, precisely under the pressure of design changes, in situations of industry conflict, when making decisions about accepting or rejecting model or cost risks (Succar, Sher, & Williams, 2013). Figure 2: Key BIM events along the project lifecycle. BIM Neural Pathways shifts the focus from an abstract list of KSAs to specific sequences of events in which these KSAs are actually needed. The event thus becomes a stimulus that activates specific competence neurons, forces them to interact in a network, and leaves a trace in the project data in the form of issues, decision logs in the CDE, QA/QC reports, or recurring problem indicators (Borkowski et al., 2023; Succar et al., 2013). This makes it possible not only to describe what competencies are needed in the BIM process, but above all which neurons are critical for a given type of event and how often they are effectively strengthened in real projects. 3.2 Typology of events and their parameterization in the BNP concept In order for events to become a tool for analyzing and designing competency paths, they must be organized into a coherent typology and described by parameters that allow them to be compared between projects. In BIM Neural Pathways, it is assumed that from the perspective of competency development, the most important events are initiating events (e.g., BIM kickoff, initial model intake), iterative events (e.g., coordination sprints, cyclical reviews of 4D/5D models), decision-making and approval events (e.g., stage freeze, approval of design variants, decisions on scope changes), transfer events (e.g., handover of models and asset data), and operational events related to the use of information in FM. Each of these types of events generates a specific configuration of information requirements and a characteristic risk profile, thereby activating a different set of competence neurons (Borkowski, 2023; Wu et al., 2021). Therefore, the key step is to parameterize the identified events. For each of them, at least several dimensions can be specified, e.g., criticality for the project outcome (e.g., impact on cost, deadline, quality of information), competency density (number and complexity of neurons that must be activated), organizational reach (number of industries and stakeholders), time pressure (degree of decision compression over time), and data trace density (type and amount of data generated in CDE and BIM tools). In project practice, these parameters can be derived from the analysis of schedules, responsibility matrices, and logs of CDE systems and coordination tools (Borkowski et al., 2023). Events encoded in this way become a kind of stimulus map for the BIM competence system. It is not a single role or position, but the structure of events in a project that determines which neural pathways will have a chance to activate and which will remain only potential. Two projects with a similar technical scope can generate radically different event profiles depending on the contract model, level of prefabrication, intensity of design changes, or degree of CDE environment integration. 3.3 Micro-loop learning embedded in a project event Against the background of the defined typology of events, the micro-loop becomes the basic unit for strengthening neural pathways. Unlike traditional courses or training, which function as external activities to the project, the micro-loop is a learning cycle that takes place during a real event. In its classic form, it consists of five consecutive phases: (i) pre-brief, (ii) project activity, (iii) capture, (iv) micro-drill, and (v) proficiency level update (Fig. 3). Figure 3: Micro-loop learning cycle embedded in a coordination sprint event. In the pre-brief phase, the team defines the intention of the event and the expected standard of performance, referring to specific competence neurons and their proficiency levels. For example, before a coordination sprint, this may mean explicitly defining expectations for neurons related to collision triage, CDE governance, and crossindustry work facilitation. At this stage, the event is also attached to the relevant part of the neuron graph, which allows for predicting in advance which competency pathways are to be strengthened (Wu et al., 2018a; Succar et al., 2013). The action phase of the project involves the normal course of the event, e.g., conducting a sprint, agreeing on changes, obtaining approval, transferring models or data. The key point is that the team's activity is monitored from the outset not only in terms of technical results, but also in terms of behaviors indicating the use of specific neurons, such as how to classify collisions, formulate issues, document decisions in CDE, or manage conflicts of interest between different industries (Borkowski et al., 2023). The capture phase consists of identifying what happened both in the data layer (issues, logs, QA/QC reports) and in the narrative layer (brief team reflection, recording of key decision moments). It is this stage that distinguishes the micro-loop from the ordinary execution of an event: instead of leaving the course of events in the participants' , evidence of the use or lack of specific neurons is selected and structured. In the micro-drill phase, the team performs short, targeted exercises aimed at accommodating knowledge or assimilating a new way of operating into existing habits (declarative and/or procedural knowledge). This may take the form of a collision triage simulation on a selected part of the model, practicing the handover procedure in a CDE test environment, or micro- negotiation scenes (Borkowski & Kubrat, 2024). The final stage of the micro-loop is updating the proficiency level for selected neurons. Unlike classic periodic assessments or certifications, here the update is based on hard evidence from a specific event, i.e., auditable proof of competence (Wu et al., 2021). For a given neuron, it is therefore possible to indicate not only the declared level of proficiency, but also the history of its reinforcement in subsequent events. In the long term, this allows us to track which neural pathways in the project have stabilized and which require further impulses in subsequent events or pilot projects (Borkowski, 2023; Succar et al., 2013). Thus, chapter three introduces the key operational concepts of BIM Neural Pathways, i.e., an event as a competency stimulus and the micro-loop embedded in it as an elementary mechanism for strengthening neural pathways. In the next step, it becomes possible to combine the event map of a specific project with the neural graph and design development paths for selected roles and teams, which paves the way for empirical validation of the BNP concept in case studies and comparative project studies. 4. BIM COMPETENCY NEURON CARD The neuron card is the basic tool for documenting a single competency in the BIM Neural Pathways model (Fig. 4). It is a kind of local neuron programming interface: it describes the purpose for which it is used, after which it is possible to learn about its activation in the behavior of team members, what traces it leaves in the project data, and how its level of mastery is measured. Unlike classic competency descriptions, which often take the form of general lists of characteristics or skills, the neuron card forces the operationalization of competencies in terms of observable behaviors and evidence, which is consistent with good practices in competency modeling in HR and occupational psychology (Armstrong, 2014; Schippmann, 2013; Spencer & Spencer, 2008). Figure 4: Example of BIM competency neuron card – “Clash triage and owner assignment”. The structure of a neuron card includes several sections that must always be filled in using the same logic so that the entire neural network is comparable. The header defines the name of the neuron in a way that is unambiguous and distinguishable from others, e.g., "Clash triage and owner assignment." The name should indicate both the type of action and the expected result. Next, the intention (purpose) of the neuron is formulated, i.e., a short answer to the question of why this unit of competence exists in the project. The intention is not a description of the activity, but of the intended effect: for example, that all relevant conflicts are classified, assigned to an owner, and linked to a project decision within a specified time frame. This distinction between goal and action is consistent with classical approaches to competencies as so-called hidden motives that manifest themselves in specific behaviors (Spencer & Spencer, 2008). The next section of the card concerns behavioral indicators. These are short, precise descriptions of behaviors that indicate that a neuron has been used in practice. They can be treated as the equivalent of 'behavioral indicators' or 'behavioral anchors' used in many competency models (Klendauer et al., 2012). The indicators do not describe personality traits, but specific actions in the context of BIM, i.e., how to formulate an issue, criteria for distinguishing critical from cosmetic collisions (triage), a decision documentation scheme, or a meeting format. The rule is quite simple: if an observer, having a neuron card at their disposal, is unable to determine whether a given behavior meets the indicator, the description is too general and needs to be clarified. The next section of the card describes the data artifacts that reflect the actions resulting from the neuron. This is a key element that distinguishes neurons from traditional competencies. According to the evidence-centered design approach, the design of an assessment system should begin with the question of what evidence will demonstrate the possession of a particular competency and where that evidence can be collected (Mislevy, Almond & Lukas, 2003). The neuron card therefore specifies whether the trace of neuron activation is found in an issue in the CDE, in the coordination tool logs, in QA/QC reports, in the model change history, or in coordination meeting minutes. Already at the stage of preparing the card, care must be taken to ensure that the evidence can be obtained automatically or, at least, is easy to audit ( ), otherwise the neuron will become difficult to measure. An important section of the card is the description of proficiency levels. The project usually uses a five-point scale from 0 to 4, but the number of levels is not as important as their description. Each level must combine three dimensions: context complexity, autonomy level, and typical behaviors, which is consistent with the logic used, among others, in the European e-Competence Framework (Ferrari & Punie, 2013). At level 0, the neuron is inactive, i.e., there is no evidence in the data. Level 1 means acting according to simple instructions, in a stable context and under close supervision. Level 2 assumes independence in typical situations, with still limited complexity. Level 3 involves working in complex, changing project conditions and the ability to engage others. Level 4 reflects the ability to design improvements to the process associated with a given neuron and to support other team members in achieving higher levels. The description of the levels should be structured in such a way that the level can be assigned based on a review of the traces in the CDE, and not solely on the basis of the employee's self-assessment. The lower part of the card specifies the prerequisites and synapses of the neuron. Prerequisites are other neurons or resources that must be present for a given competency to manifest itself. They can relate to technical skills, process knowledge, or formal authorizations. Stabilizing and co-activating synapses are relationships with other neurons that usually activate simultaneously or help maintain the durability of the competency effect. For example, the neuron responsible for collision triage may be stabilized by the inter-industry conflict facilitation neuron or the CDE standards management neuron. In practice, these relationships help to design development interventions: instead of raising the level of a single competency in isolation, a set of micro-loops is designed to strengthen the entire network segment. This is consistent with observations that effective competency models should reflect the links between behaviors rather than treating each competency as an independent trait. The very process of creating a neuron chart combines elements of classic competency modeling, job analysis, and evidence-based assessment design. It usually begins with the selection of a specific project event in which a given competency is crucial. This is followed by short interviews or workshops with practitioners who achieve aboveaverage results in this event. This type of approach, which draws on the analysis of the behaviors of top performers, is widely used in research on professional competencies. During the workshop, participants describe step by step what they do in typical and difficult situations, and the facilitator identifies patterns of behavior that distinguish experts from average performers. The second stage is to translate these patterns into behavioral indicators and determine what data artifacts are associated with them. At this point, evidence-centered design logic is applied: for each potential indicator, questions are asked about what evidence would confirm its occurrence, where it can be found, and how it can be recorded in the course of normal work without overburdening the team. If no realistic data source can be identified for a given behavior, the indicator is modified or rejected. The third stage is the design of proficiency levels. Here, inspiration can be drawn from ready-made frameworks such as e-CF, which show how to describe increasing complexity of context, autonomy, and impact, while avoiding purely quantitative counting of years of experience. The project team writes out scenarios for the actions of an expert, an intermediate, and a beginner, and then distills descriptions of the levels from them. A useful technique is to formulate short stories (storytelling) illustrating typical behaviors at each level, and only later shorten them to concise descriptions in the neuron card. The fourth stage involves integrating the neuron with the network of other competency units. At this stage, prerequisites and synapses are identified using both expert knowledge and analysis of co-occurrence of traces in the data: if certain artifacts almost always appear together, this may indicate connections between neurons. This approach is similar to building a competency architecture at the organizational level, as described by the authors of competency models for digital and ICT (Information and Communication Technology) professions (GarcíaBarriocanal et al., 2012). The final step in creating the card is its validation in a pilot project. This involves the team using the card during micro-loops for a specified period of time: in the pre-brief phase, it refers to indicators and levels; in the capture phase, it checks whether the indicated artifacts are actually being created; and in the proficiency update phase, it attempts to assign a level based on the evidence collected. If the card proves to be too general, inconsistent, or difficult to use, it is modified iteratively. This cyclical nature of card refinement fits well with the literature on microlearning and short, work-embedded development interventions, which are more effective than one-off training sessions detached from context (Leong et al., 2020). As a result, the neuron card becomes not only a description of competencies, but also an operational contract between the project team, the HR (Human Resources) department, and the organization's IT systems, e.g., ERP (Enterprise Resource Planning). It specifies exactly what is to be strengthened in subsequent micro-loops, how it will be measured, and what other neurons are linked to this process. This makes it possible to gradually build a coherent BIM competency graph, which is not an abstract list of skills, but a network of related units of knowledge and practice, anchored in real project data. 5. APPLICATION OF THE NEURAL PATHWAYS BIM MODEL IN A PILOT PROJECT 5.1 Constructing an "event-neuron" matrix The starting point is to organize projects according to repeatable events, not just phases or milestones. In practice, existing information and project management frameworks are used, such as ISO 19650-1 (2018), which describes key information processes in the life cycle of a building, and PMBOK® (2021), which organizes a project through processes, process groups, and knowledge areas. By analyzing the schedule, information management plan, and responsibility matrices (RACI), the team identifies a set of BIM events, e.g., BIM kickoff, model intake, coordination sprints, stage freezes, 4D/5D baseline, model and asset data handover. Then, for each of these events, a list of expected information results is created (e.g., agreed discipline models, list of accepted collisions, updated model risk matrix) and the neurons that must be activated according to their cards in order for the result to be achieved are identified. This is a direct application of the evidence-centered design logic mentioned earlier, so instead of asking generally "what competencies are important in this event," an argument is constructed along the lines of "if the event ended with result X, then we should see evidence of the activation of neurons A, B, and C in the data" (Mislevy, Almond, & Lukas, 2004). The set of events (rows) and neurons (columns) forms an "event-neuron" matrix (Fig. 5). Each cell specifies the degree of neuron activation in a given event, e.g., on a four-point scale: 0 – no connection, 1 – auxiliary, 2 – significant, 3 – critical. In practice, this scale should be established in workshops with key project stakeholders so that it reflects both the technical perspective (e.g., BIM coordinators) and the organizational perspective (e.g., contract managers). Once the matrix has been completed, a heat map of the project is created, showing which events have the highest competency density and which neurons are critical at many points in the process.