[This page covers needs, drives, motivation, and affect in relation to emotion]
Emotion is something everyone experiences but it remains somewhat difficult to define precisely, in part because of our strong subjective experiences (Ekman & Davidson, 1994; LeDoux, 2000). From the computational and neuroscience perspective taken here, we can provide a more precise definition:
Emotion reflects the activity of midbrain and brainstem systems that have been shaped by evolution to guide the behavior of the organism in adaptive ways to satisfy its needs in relation to the external and internal state. Most emotional states involve midbrain neuromodulatory systems such as dopamine and serotonin that broadcast signals reflecting important changes in body state and have widespread effects on the entire brain.
These brain systems have traditionally been categorized as the limbic system (but this is not very precise as discussed there), and are anchored by the hypothalamus, periaqueductal gray, parabrachial area, and associated nuclei that directly sense the internal body state (interoception), which is conveyed in part by the vagus nerve via the medulla oblongota.
Multiple layers of additional processing and control over the brainstem systems have been added and elaborated over the course of brain evolution, including:
• The amygdala, which makes connections between sensory inputs processed by the cortex (visual, auditory) and underlying body states.
• Ventral and medial basal ganglia (e.g., the nucleus accumbens) including the ventral pallidum output pathway that interconnects widely with the midbrain areas.
• The lateral habenula, which integrates a wide range of sensory inputs to determine when to “give up”, according to the Rubicon framework.
• Ventral and medial prefrontal cortex areas including the insula, that receive and send extensive connections to all of the other emotion-processing areas, providing a higher level of integration and control that is the focus of the Rubicon goal-driven framework.
• Ventral hippocampus, which is also extensively interconnected with all of these areas, and provides the ability to rapidly encode emotional states associated with distinct events in time and space.
Psychologically based accounts have struggled with attempts to categorize the number and nature of emotional states, and to delineate boundaries between emotion and cognition (Ekman & Davidson, 1994; LeDoux, 2000). By contrast, the Rubicon framework is based on the idea that the goal-driven, motivational aspect of emotional states is essential for driving the learning and cognitive control that shapes all cognitive functions.
Computationally, much of this falls within the domain of reinforcement learning (RL). However, RL typically only deals with a very limited scope of “emotion”, as captured in a single scalar reward value, whereas the Rubicon framework encompasses a broader range of states and the relationship between needs / drives and the current internal state, including goals and motivational state.
The term affect typically encompasses a broader scope than emotion, but we use them essentially interchangably here, under the above definition. Likewise, the psychological distinction between mood and emotion is not a primary concern in our framework, where the relevant internal states can extend across a range of different timescales.
Categorizing emotion
Figure 1:
Valence (positive vs. negative) vs. arousal (high vs. low activation) in the 2D circumplex model of emotion.
Figure 2:
Plutchik’s (2001) wheel of emotions, with arousal (intensity) represented as distance from the center along any of 8 different categories of opponent emotions.
From a subjective, psychological perspective, emotions can be categorized most broadly in terms of valence (positive vs. negative) and arousal (high vs. low) (Figure 1). Progressively more differentiated systems have been developed, with Plutchik’s wheel of emotions providing a nice comprehensive set (Figure 2; Plutchik, 2001).
Figure 3:
Maslow’s hierarchy of needs, progressing from most essential to those that gain salience once the lower-level needs are satisfied.
What about more basic feelings, like hunger, thirst, etc? Maslow (1943) developed a hierarchical schema to organize human needs, with the most basic ones providing the foundation, and higher-level ones only entertained once the basic needs are satisfied (Figure 3).
The evolutionarily most ancient functions of the hypothalamus can be categorized into the five Fs: feeding, foraging, fleeing, fighting, and fornication, which comprise the most important survival-relevant behaviors managed by the brain, and have direct hormonal correlates. In this framing, most people spend the majority of waking time in the foraging mode (with information replacing nutrients), which is where goal-driven behavior is most relevant.
Figure 4:
Emotional states as providing guidance toward satisfying needs, with a lack of need satisfaction generally driving negative emotional states, while need satisfaction drives positive ones. The list of needs is ordered with the most essential needs at the bottom, as in Maslow’s hierarchy. We don’t usually think of things like hunger and thirst as emotions, but according to this system, they play the same overall role, and share many neural substrates in common. Resources refers to any kind of material thing needed to survive in the current physical environment (money, territory, nesting material, tools, building materials, etc). The social factors include S = self perspective and O = other perspective. Most of these needs apply across all species of mammals and other vertibrates to variable extents. The social needs are particularly important for shaping flexible, open-ended human cognition.
We can build on this schema to define a comprehensive organization of emotions organized around the things that humans need to do from an evolutionary perspective, as summarized in Figure 4. This table provides a place for most of the terms in Plutchik’s wheel of emotions, and covers most of the needs in Maslow’s hierarchy, providing a reasonable basis for seeing how emotions help guide us toward the things we need to survive and thrive.
Interestingly, most of these needs apply across all species of mammals and even more broadly to all animals. This suggests that emotion per se is widely shared across the animal kingdom, not something that is uniquely human, which is consistent with emotion being rooted in the deepest, most evolutionarily ancient portions of the brainstem. What the human brain adds is a much enlarged neocortex and an associated ability to gain conscious awareness of all this emotional stuff going on, largely via the extensive connectivity from these brainstem areas into the medial and ventral regions of the prefrontal cortex.
Figure 5:
Drive reduction theory according to Hull, 1943. Basic needs create drives when those needs are not satisfied, and behavior is then recruited to satisfy those drives.
From a computational perspective, we can define each of these needs as having a current drive level (Hull, 1943), which reflects the current lack of satisfaction of the need (Figure 5). This is the model incorporated in the Rubicon model, with a variable number of need factors that each have associated drives and USs that satisfy them.
Social needs
Like many other species, humans are strongly social animals, and we depend on others to survive and thrive. There is a strong correlation between brain size and size of social networks in primates (Dunbar, 1992; Dunbar, 2016), suggesting that we owe our big brains to being highly social animals: it takes careful thought and planning to navigate the complexities of the social world.
Furthermore, Tomasello (2001) argues that humans have a unique drive to share that is not evident in even our closest primate relatives (e.g., chimpanzees). This sharing instinct is what drives the pervasive nature of our cultural evolution, where we acquire much of our knowledge from the accumulated wisdom of those who have come before us.
A major function of emotion is to communicate and share our internal states with others, so they can help us satisfy our needs as well. This is nicely conveyed in the movie Inside Out with respect to the important role of sadness. Considerable work has identified a set of facial expressions that are universally recognized across cultures (Ekman & Friesen, 1975): anger, disgust, fear, happiness, sadness and surprise. Note that this small set is not thought to be exhaustive, and not all emotional states are communicated: the ones on this list are those that are clearly useful to communicate.