What Is Anxiety?

Last week’s post about positive and negative emotions led me to ask what emotions are and how they arise in the brain. The neuroscientiest Joseph LeDoux’s excellent book, Anxious: Using the Brain to Understand and Treat Fear and Anxiety (2015) lays out what we know about anxiety. Two of his points challenged concepts I have relied on for years: the idea of unconscious anxiety, and that it is the brain’s basic response to perceived threats.

LeDoux insists on defining anxiety as a feeling that occurs in consciousness. This is not semantic hair-splitting—it starts from his rigorous insistence on specifying how we know what we know. Limiting the concept of anxiety to what can be put into words turns out to be useful, for it leads to an exploration of the nonconscious brain events which contribute to the feeling of anxiety but are clearly not themselves anxiety. (I will use LeDoux’s “nonconscious” rather than the psychoanalytic term “unconscious” when discussing brain mechanisms.)

I had trouble with this restricted definition. Like most mental health professionals, I have seen many patients who denied their anxiety or dealt with it counterphobically. I studied how heart surgery patients handle anxiety for my M.D. thesis, and I still recall preoperative patients whose sweating faces and rigid postures communicated high levels of anxiety even though they insisted they felt fine about their surgeries. Other emotions can be denied or suppressed, though I balked when a psychotherapy supervisor suggested a patient was experiencing “masked depression.” “How could I tell if it’s there,” I asked, “if it’s masked?” That is one of LeDoux’s points—we are on thin ice when we talk about supposed phenomena we have no way to validate or monitor.

Anxiety which, to use psychoanalytic language, the ego manages with denial or reaction formation is still, as I understand it, available to consciousness on some level in a way the nonconscious components of anxiety LeDoux studies are not. But it is important to recognize that our neuroscientific understanding of consciousness is quite limited, and that while there have been neurobiological studies of dissociation and repression, we are not close to explaining mental operations like the ego mechanisms of defense.

Charles Darwin drew on observations of animal behavior and humans from different cultures to argue that emotions like pleasure, grief, anger, and fear are expressions of evolved mental states we share with animals. The idea is that evolution constructed our brains to respond to external threats with basic emotions like fear, which are then expressed in physiological changes like palpitations, behaviors such as freezing, and thoughts of “I’m scared.” Today’s popular concepts of anxiety have built on Darwin’s view of anxiety as a basic emotion elicited by external threats.

LeDoux challenges the notion that anxiety is a basic state of the brain. Recounting experiments in animals and humans, he shows that there is no structure or circuit that produces a central emotion of anxiety. A threat triggers many distinct phenomena, some of which have been only partially elucidated. In the relatively simple case of conditioned responses, he identifies six behavioral components, each of which has specific circuitry involving several areas of the brain. We would not identify any of these as the feeling of anxiety or fear. They are: increased attention to threats (hypervigilance), impaired ability to discriminate threat from safety, increased avoidance, heightened reactivity to threat uncertainty, overvaluation of threat significance and likelihood, and maladaptive behavioral and cognitive control in the presence of threats. Hypervigilance, for example, begins in the brain with excessive activation of the amygdala, which signals the periaqueductal gray area to initiate defensive behavior responses such as freezing. Arousal systems in the basal forebrain and brain stem are also activated, facilitating processing in the amygdala and sensory areas of the cortex. Another example is avoidance, which involves the amygdala, nucleus accumbens, and dorsal striatum, as well as insular, orbitofrontal, and cingulate cortical areas.

Thus something as apparently simple as Pavlovian conditioning, turns out to be more complex than I am able to imagine. It leads me to wonder what must be going on in the panicked brain of an agoraphobic patient at a crowded mall who is separated from the companion he clings to, a fairly common problem in clinical psychiatry.

But where in these brain systems is the anxiety? It is clearly not simply the activation of the amygdala, which some call the brain’s “fear center.” And it is not excessive activation of the brain as a whole: while generalized arousal accompanies anxiety, many circuits other than arousal are involved. Nor, LeDoux argues, is the feeling of anxiety simply the sum of all these smaller events. To feel anxiety requires some level of conscious awareness.

Consciousness is a great mysteries which science has not yet elucidated. (The other is the origin of the universe.) LeDoux moves us forward here, and this paragraph and the next are a gross abridgement of his detailed account, which, I am sure he would agree, is itself far from complete. Consciousness requires working memory and attention. Central for memory is the medial temporal lobe memory system, including the hippocampus and surrounding cortical areas such as the entorhinal cortex, parahippocampal cortex, and perirhinal cortex. These have reciprocal connections to the parietal, temporal, and frontal lobes. Active, conscious memory also involves the prefrontal cortex. Attention involves prefrontal and parietal lobe circuits and controls what is presented in working memory.

LeDoux uses a cooking analogy to summarize our present understanding of anxiety. If the feeling of anxiety is a soup, working memory is the pot. The ingredients, he lists, are sensory processing; survival circuit activity; brain arousal, bodily response feedback; memory; and the executive functions of attention, labeling, monitoring, and attributing. This metaphor is satisfying in that it includes many elements we recognize as important, but, like the biopsychosocial model I discussed in my post on conceptual models, is more a list than an explanation. It does not tell us how the ingredients interact to generate the conscious experience of anxiety, and it is limited in its power to generate clinical hypotheses.

I am left with a feeling of discouragement: if neuroscience can explain so little about something as fundamental to psychiatry as anxiety, aren’t clinicians better off continuing to base practice on traditional diagnoses, concepts, and theories? These have at least stood the test of time. Immersion in emerging neuroscience could be paralyzingly complex with limited clinical yield. And, unless we are as rigorous as LeDoux, we run the risk of adopting simplistic biological concepts which could mislead us with possibly tragic results. (I think of the harm from thousands of lobotomies in the 1940’s and 1950’s, performed to disconnect the prefrontal cortex from the rest of the brain.)

But as I have previously discussed, for psychiatry, the tried and true has serious limitations. And even if present neuroscience is both too complex and too incomplete to provide much guidance, it at least raises questions which may be useful in clinical situations. First, we must recognize that emotions such as anxiety are complex brain phenomena which are both difficult to describe and suffused with nuance from the patient’s memories and current mental state. What a patient means by anxiety may be very different than what the clinician understands. Our techniques for increasing precision—having the patient describe in detail a particular situation when she experienced anxiety, asking where in her body she feels it, and respecting her need to use metaphors when other language is inadequate—are still useful.

With anxiety in particular, it is helpful to break it down, if not into its fundamental components, at least into types. Both the Beck and Zung anxiety scales include items about generalized anxiety (worry), panic, and arousal. It is helpful to separate these so treatment can be targeted. Scales based on current DSM diagnoses, such as the GAD-7, the Liebowitz Social Anxiety Scale, and the Panic Disorder Severity Scale, may be more helpful.

I expect to discuss treatments for anxiety in another post, but for now it is worth noting that two partially effective treatments for anxiety, meditation and benzodiazepines, exert some of their effects by dampening overall brain arousal. (LeDoux notes, however, that benzodiazepines, and probably meditation, have more specific effects as well.) We can hypothesize that they may be more effective in patients whose level of general arousal is high.

This has been an interesting journey for me. An important psychiatric concept—anxiety—which has been central to my understanding and treatment of patients, turns out to be much more neurobiologically complex than I imagined. It leaves me with a feeling of humility about how little we know, but such humility may itself be useful.

Further reading: Joseph LeDoux, Anxious: Using the Brain to Understand and Treat Fear and Anxiety, 2015.


Edvard Munch, The Scream.

Recent Posts