The Squeeze in the Middle: Why Middle Managers Are the Most Neurobiologically Strained Group in Organizations

 KI forecasts are increasing the pressure on middle management. But the real problem is structural in nature - and can be precisely described from a neuroscientific perspective. What organizations must do now.

Caught in the middle: The middle manager in the AI transformation, caught between top management and employees. Photo generated by AI

In leadership research, middle management is regularly described as the “translation layer” between strategy and execution. What this description overlooks: this layer operates under conditions that - neurobiologically speaking - are among the most demanding in the entire organization. Not because the individuals in these roles are particularly vulnerable. But because the structural position they occupy simultaneously puts pressure on multiple fundamental human need systems.

This article develops a precise analysis of this multi-layered strain, connects it with current empirical data and AI forecasts, and derives implications for strategic leadership and organizational design.

The Empirical Starting Point: What the Data Shows

The Global Leadership Forecast 2025 by DDI (10,796 leaders, 2,185 HR professionals, 2,014 organizations, 50+ countries, 24 industries) is the most comprehensive longitudinal study available on leadership behavior. Its findings for middle management are clear:

71 percent of all surveyed leaders report significantly increased stress - eight percentage points more than in 2022. According to DDI, frontline managers - the lowest leadership level with direct team contact - are three times more concerned about the impact of AI on their leadership role than those at the C-level. Trust in direct supervisors declined from 46 to 29 percent over the same period - a drop of 17 percentage points.

These figures are significant in themselves. However, what the DDI report does not provide - and this is a crucial distinction - is a neurobiological explanation for these developments. The following interpretation is analytical: the increased strain on frontline managers can - complementing the DDI empirical data - be explained using the SCARF model, because the sandwich position structurally activates multiple social threat domains simultaneously. This connection is the author’s interpretation, not a statement from the DDI study.

The SCARF Model: A Neurobiological Analysis of the Sandwich Position

The SCARF model, developed by David Rock (NeuroLeadership Institute, 2008), describes five social domains in which the brain evaluates a situation as either safe or threatening: Status, Certainty, Autonomy, Relatedness, and Fairness. The model is a heuristic framework that translates empirical findings from social neuroscience into practically applicable language.

The neurobiological basis is well established: Eisenberger, Lieberman, and Williams (UCLA, Science, 2003) showed in an fMRI study that social exclusion and rejection activate the same brain regions as physical pain - the anterior cingulate cortex. Social threat is not an abstract category for the brain. It activates the same survival system as physical danger.

Middle management does not experience diffuse overload. It experiences a precisely structured neurobiological threat response - in all five SCARF domains simultaneously.

Applied to middle management, the picture is as follows:

Status: Excluded from the strategic level, increasingly critically evaluated by the team level. Threat from two directions simultaneously.

Certainty: AI forecasts call parts of one’s role into question - without a clear time horizon. The brain reacts more strongly to unpredictable threats than to defined ones.

Autonomy: Responsibility for execution without co-creating direction. Accountability without decision-making autonomy generates sustained autonomy stress.

Relatedness: Positioned between two groups without full belonging to either. Social isolation as a structural feature of the role.

Fairness: High performance expectations, asymmetric recognition. The brain registers imbalances in social exchange as threats.

All five domains under pressure simultaneously - this is the neurobiological signature of the sandwich position. The physiological consequences of sustained social threat perception are well documented in stress research: the sympathetic nervous system is activated, cortisol levels rise, and activity in the prefrontal cortex - responsible for reflective thinking, empathy, and strategic judgment - declines. These relationships are part of established stress physiology and are applied here as an analytical framework to the described role situation - a direct empirical measurement for middle managers does not exist.

AI and Middle Management: What Gartner Actually Says

The discussion around AI and middle management is shaped by a Gartner forecast from October 2024: by 2026, 20 percent of organizations will use AI to flatten hierarchies and, in doing so, reduce more than half of middle management positions.

This forecast is a strategic scenario - not a description of developments that have already occurred. Gartner itself emphasizes in more recent studies that 72 percent of surveyed CIOs are not reaching the break-even point with their AI investments. And the latest Gartner forecast for 2026 warns of an opposite risk: atrophy of critical thinking due to excessive AI use will lead 50 percent of organizations to introduce “AI-free” competency assessments.

The Gartner forecast is a scenario, not a certainty. And it is relativized by Gartner’s own newer data. The strategic fallacy: automation of tasks is not the same as elimination of roles.

What is actually under automation pressure: reporting, performance monitoring, standardized email reporting, coordination of routine processes. What is not substitutable: situational judgment, emotional intelligence, trust-building through direct interaction, navigating ambiguity in human systems. These competencies are not replaced by AI - they become more relevant as AI relieves routine tasks.

The Transmission Dynamic: Why This Is a System Problem

One of the most underestimated phenomena in leadership research is the transferability of emotional states from leaders to teams. The phenomenon of emotional contagion is well documented: Hatfield, Cacioppo, and Rapson (1993) showed that moods and affective states are transmitted between people through facial expressions and nonverbal signals. Later leadership research - among others, Sy, Côté, and Saavedra (2005, Journal of Applied Psychology) - demonstrated this effect specifically in leadership contexts: the affective mood of a leader measurably transfers to team members and influences their performance.

The practical consequence: research on emotional exhaustion among leaders shows that leaders under sustained strain are more likely to fall into a laissez-faire leadership style, which demonstrably impairs psychological safety in teams. The ability to regulate, orient, and stabilize is limited by sustained exhaustion - not necessarily completely eliminated. This is a plausible explanation for the decline in trust in direct supervisors: not necessarily worse leaders, but leaders operating under deteriorating structural conditions. This causal interpretation is analytical - the DDI study demonstrates the decline but does not identify a neurobiological cause.

Four Strategic Fields of Action for Organizations

1. Understand the Sandwich Position as a Structural Design Problem

The first step is conceptual. In many organizations, the strain on middle management is treated as an individual resilience problem: the leader should “manage stress better,” “delegate more,” or “prioritize more clearly.” This falls short.

The described multi-layered strain is not a symptom of personal inadequacy - it is the neurobiologically predictable response to structural conditions. Organizations that understand this shift the logic of responsibility: from “What must the leader change?” to “What conditions do we create under which good leadership is possible?”

Concretely, this means: explicit discussion at executive level of the conditions middle managers face. Regular, structured conversations with this level - not as a control mechanism, but as a signal: this perspective matters. And a willingness to make structural adjustments when conditions systematically undermine leadership quality.

2. Systematically Improve Communication About Uncertainty

The brain reacts more strongly to unpredictable threats than to defined ones. This means: uncertainty about the future of one’s role is neurobiologically more stressful than a clear negative statement. Organizations that communicate internally what they know about AI developments - and what they do not - what task areas will change and which will not, and what development paths exist for middle managers, significantly reduce perceived threat - regardless of the content of the message.

3. Position AI Enablement as an Investment in the Middle Layer

The most productive response to the AI debate in middle management is not reassurance (“AI will not make you redundant”), but enablement. Leaders who learn to use AI tools as assistants - for drafts, research, structuring, critical reflection - gain cognitive capacity for tasks that require human judgment. This is a concrete transformation of the role: from coordination and reporting to judgment and human leadership.

4. Anchor Self-Regulation as an Organizational Capability

Self-regulation - the ability to recognize and manage one’s own nervous system state - is not only a personal resilience resource for middle managers. It is a leadership-relevant competency that directly transfers to teams. Organizations that explicitly embed self-regulation in leadership competency models, provide coaching and development formats, and culturally signal that leaders are allowed to speak about their own strain create the foundation for a more stable middle management layer.

Conclusion: Asking a Different Question

The discussion about middle management usually revolves around competence, efficiency, and AI substitutability. This discussion leads to a dead end because it ignores the key variable: the structural conditions under which this role is performed.

The neurobiological analysis shows: middle management operates in a systematically induced multi-stress state that limits cognitive flexibility, empathy, and judgment - precisely the capacities required for transformation, change management, and team leadership. Those who want to strengthen this level must strengthen the conditions under which it operates.

The most productive question is not: “How do we develop middle management?” It is: “How do we design the conditions under which this level can realize its potential?” That is a different - and strategically more relevant - question.

About the Author

Dominique Giger is a transformation expert, coach, and speaker with a Master of Science in Computer Science from ETH Zurich. She brings more than 18 years of international experience in transformation and change projects, guiding organizations and leaders toward resilient, high-performing work cultures.

Her work combines neuroscientific insights with practical leadership experience, with a particular focus on mental strength, sustainable performance, and authentic leadership in complex work environments.

In her podcast "Y-SHIFT: The Next-Level Mindset & Transformation Podcast", she regularly shares insights from the worlds of modern psychology, neuroscience, and leadership development.

More on this topic in Episode 33 of the podcast: Between Pressure from Above, Below, and AI - Why Many Are at Their Limit  (in German)

YouTube: https://youtu.be/fIMgvU92PLk

Spotify: https://open.spotify.com/episode/7uLB946OfSMTima7TTybT6

Apple Podcast: https://podcasts.apple.com/us/podcast/folge-33-zwischen-druck-von-oben-unten-und-ki-warum/id1801021329?i=1000763396298

References

• DDI (2025). Global Leadership Forecast 2025

• Rock, D. (2008). SCARF: A brain-based model for collaborating with and influencing others. NeuroLeadership Journal, 1(1), 44-52.

• Eisenberger, N. I., Lieberman, M. D., & Williams, K. D. (2003). Does rejection hurt? An fMRI study of social exclusion. Science, 302(5643), 290-292.

• Eisenberger, N. I. (2012). The pain of social disconnection. Nature Reviews Neuroscience, 13, 421-434.

• Hatfield, E., Cacioppo, J. T., & Rapson, R. L. (1993). Emotional contagion. Current Directions in Psychological Science, 2(3), 96-99. https://doi.org/10.1111/1467-8721.ep10770953  |  Sy, T., Côté, S., & Saavedra, R. (2005). The contagious leader: Impact of the leader’s mood on the mood of group members, group affective tone, and group processes. Journal of Applied Psychology, 90(2), 295-305.

• Desrumaux, P. et al. (2024). The ripple effect of strain in times of change: how manager emotional exhaustion affects team psychological safety and readiness to change. Frontiers in Psychology / PMC.

• Gartner (Oktober 2024). Gartner Unveils Top Predictions for IT Organizations and Users in 2025 and Beyond.

• Gartner (Oktober 2025). Gartner Survey Finds All IT Work Will Involve AI by 2030.

• Gartner (Oktober 2025). Gartner Unveils Top Predictions for IT Organizations and Users in 2026 and Beyond.

• Bandura, A. (1997). Self-efficacy: The exercise of control. W. H. Freeman.

• DDI (2025, April). Only 19% of Rising Leaders Have Delegation Skills.