salmon logo
"Anxiety is a mental disorder when it is more intense than is justified by the actual threat" (Lickey and Gordon, 1991)

Visit the SALMON Bookshop for recommended books on this topic  

Benzodiazepine Drugs & Anxiety
Author Paul Kenyon

Anti-anxiety drugs

Serotonin and anxiety

GABA and anxiety

This lecture builds on an earlier lecture on anxiety. It begins with some background information on the efficacy and use of anxiolytic drugs. Some patients on these drugs experience side effects, therefore there is still a need for research in this area to develop safe and effective medicines.

The heart of the lecture explores Stein's theory of anxiety which argues that

  • the acute sedative effects of benzodiazepine (BZP) anti-anxiety drugs are produced by changes in the noradrenergic (NA) system,
  • changes in serotonin (5-HT) metabolism account for the chronic anxiety-reducing 'anxiolytic' effects of the drugs.

We then explore how this theory was tested by using various drugs that affect 5-HT. We examine an important experiment by Kilts et al, (1982) that fails to support the theory - in fact they find results opposite to those predicted by Stein.

Finally we turn our attention to the relationship between anxiolytics and the neurotransmitter GABA. I leave it up to you to decide how well a BZP/GABA interaction can explain the clinical effectiveness of BZPs, and how you would set out to find a novel anxiolytic.

Learning objectives
After studying the material on this page you should be able to:
  • Recognize the psychological and physiological symptoms of anxiety
  • Describe the reasons for the change in benzodiazepine prescription habits during the last 40 years
  • Compare and contrast the therapeutic and adverse effects of barbiturate and newer anti-anxiety drugs
  • Describe the components of the Geller-Seifter paradigm
  • Describe the effects of benzodiazepines on the Geller-Seifter paradigm
  • Describe, and explain the significance of, the correlation between the clinical potency of benzodiazepines in humans, and their effects on the Geller-Seifter paradigm
  • Describe intracranial self stimulation and explain how it can be used to study the biological basis of behaviour
  • Describe the roles of noradrenalin and serotonin in reward and punishment
  • Describe Stein's theory of reward & punishment
  • Describe Stein's explanation for the anti-anxiety and sedative effects of benzodiazepine drugs
  • Compare and contrast the acute and chronic effects of benzodiazepines on Geller-Seifter performance
  • Evaluate Stein's theory of the role of serotonin in anxiety with reference to the clinical effects of buspirone, and the effects of methysergide, PCPA, and 5-HTP on conflict behaviour in animals
  • Describe the interaction of benzodiazepines with the GABA receptor
  • Evaluate the evidence for the view that benzodiazepines reduce anxiety by interacting with the GABA receptor

The nature of anxiety

Homer tries to train Bart to get over his fear of falling, Video clip available here

Lisa experiences anxiety when babysitting Maggie. Video clip available here

Sometimes the terms 'fear' and 'anxiety' are used interchangeably, or as synonyms.  However for the sake of simplicity it could be argued that:

  • Fear is the short-term response produced by stress. The  amygdala may be a control centre for fear, receiving fear-related sensory information and transmitting fear-related motor instructions. Time magazine has produced a very nice animation -  Inside the Anxious Brain - which explains the role of the thalamus, amygdala and hippocampus in fear.
  • Anxiety has similar symptoms to fear, but persists after the stress has ended..

Craig et al (2000) discuss the relationship between fear and anxiety in great depth.

Anxiety becomes a problem when it persists too long beyond the immediate threat. Sometimes there's an obvious cause, as with the shell-shocked soldiers of World War I but for many of us there is no obvious reason for our anxiety. About 10% of people experience anxiety at some point in their lives.

See Lickey & Gordon (1991) for a fuller description of  the symptoms of anxiety

Anxiety is an unpleasant state accompanied by:
    • apprehension
    • worry
    • fear
    • nervousness

Increased autonomic activity leads to:

    • increase in blood pressure
    • increase in heart rate
    • erratic respiratory rate
    • decreased salivary flow leading to dry mouth & throat
    • gastrointestinal disturbances

Benzodiazepine prescriptions

Benzodiazepines are used to treat a variety of disorders, including anxiety, panic attacks, insomnia and muscle spasm.

From 1965 to 1985, 1.5 billion prescriptions were written in the US for benzodiazepines. (Lickey & Gordon, 1991) .

In 1997, 18 million prescriptions for benzodiazepines were written in the UK .

After a peak in the mid-1970s there was a decline in the use of these drugs because of fears that they could lead to physical and psychological dependence.

Because of this, the UK The Committee on Safety of Medicines issued the following advice:

A study published in The Lancet (24 October 1998) highlighted a significant association between taking benzodiazepines, and the risk of a road traffic accident.

Efficacy of anxiolytics

This diagram compares anxiolytic efficacy in chronic anxiety patients treated with diazepam, phenobarbital and placebo.

Patients were assessed two and four weeks after beginning drug treatment.

The reduction in anxiety was measured on a scale of 0 to 3, where 0 = no improvement, 3 = very much improvement.

The results show that diazepam and phenobarbital are not consistently different in relieving symptoms.

At one time barbiturates such as phenobarbital were the only effective drug treatments for anxiety but they are no longer used because they can be addictive and overdosing is fatal.

Side effects of anxiolytics

This diagram shows the side effects of diazepam, phenobarbital and placebo after 2 and 4 weeks of treatment.

A significant number of patients taking phenobarbital reported unpleasant side effects.

Patients taking diazepam were less likely to experience drowsiness or dizziness than those treated with phenobarbital.

There was no significant difference in reported side effects between benzodiazepine and placebo treated patients.

Nevertheless, benzodiazepines do cause side effects in some patients, and these drugs can be addictive and may interfere with complex sensory-motor tasks such as driving

Adverse effects of benzodiazepines
Lack of tolerance produces :
  • sedation
  • memory impairment
  • lack of concentration
  • motor un-coordination
  • muscle weakness
  • acute confusional state
Physical dependence causes rebound withdrawal effects which include:
  • insomnia
  • anxiety
  • apprehension
  • irritability
  • palpitations, tremor, vertigo, sweating

Long-term use of benzodiazepines is now avoided, 15 to 20% of patients on anti-anxiety drugs for more than one year report problems. Therefore there is a need for new anxiolytic drugs with fewer adverse side effects.

The need for new anti-anxiety drugs
The search for new drugs to treat anxiety starts with laboratory studies using animals in tests that are designed to model human anxiety. One type of test involves placing the animal in a ' conflict ' situation. These tests are based on the idea that ' approach-avoidance ' situations evoke anxiety. For example, we experience conflict if we have to choose between eating a very attractive dessert (approach) and avoiding an increase in our body weight (avoidance).

The Geller-Seifter paradigm is a conflict test in which rats are trained to bar press to receive sweetened milk (a very desirable reinforcement for a rat). There are two schedules of reinforcement in operation.

Conflict is introduced into this situation by delivering electric shock to the animal's feet every time it bar presses during the CRF schedule.

Effect of benzodiazepines in Geller-Seifter paradigm

This diagram shows the performance of a rat in the Geller-Seifter paradigm after a saline injection (upper panel) and following the injection of a benzodiazepine .

The effect of the benzodiazepine is revealed in performance on the CRF schedule of reinforcement where the rat is presented with a conflict between bar-pressing for a guaranteed reward, but at the cost of an inevitable electric shock to its feet

Important points to note:

  • the drug does not affect performance on the VI component
  • performance on the CRF (punished) schedule is increased by the anti-anxiety drug
    Schedule of reinforcement
    VI (unpunished) CRF (punished)
Injection Saline high, stable response rate no responding
Benzodiazepine no change to high, stable response rate increased response rate

Refer to an earlier lecture on anxiety for discussion of the specificity of this drug effect

Explanation of cumulative record

IIf you are unsure how a to interpret a cumulative record run the animation to see how a cumulative recorder represents the rate of rat's lever pressing.

The slope of the tracing made by the response marking pen is a function of how many responses are made in a fixed time period. The greater the response rate the steeper the slope. A flat tracing indicates that the rat has not pressed the lever in the Skinner (operant) box.

The animal injected with saline shows a flat tracing during the CRF component in the diagram above.

In contrast response rate was high, and there is relatively little variability in response rate under drug and saline conditions under the VI schedule of reinforcement.

Correlation of rat punishment potency with clinical potency
Correlation of rat punishment potency with clinical potency The term 'benzodiazepine' refers to a family of drugs that share a common chemical structure.

There is a very strong correlation ( r=.987 ) between the dose of benzodiazepine drug that needs to be given to a patient to control their anxiety, and the dose of the same drug that reduces conflict in animals.

In this diagram

This suggests that there may be a common biochemical basis for the effect of the drugs in humans and animals. One of the first theories to explore this idea was Stein's theory that the anti-anxiety effect of benzodiazepine drugs involved the serotoninergic system in the brain.

Apparatus used to study intracranial self-stimulation
lever triggers a rewarding electric stimulus Stein used a technique called intracranial self stimulation to investigate the role of noradrenalin and serotonin in punishment and reward .

It has been known for some time that rats will press a lever in order to have electrical stimulation delivered directly to specific areas of the brain - so called 'reward areas'. In addition they will press a lever in order to turn off stimulation to 'punishment areas' of the brain.

This diagram shows a rat pressing a lever that controls the delivery of electrical stimulation to its brain.

Stein was interested in the chemicals that controlled reward and punishment systems in the brain. To study this question he injected neurotransmitter substances directly into the brain and studied their effects on intracranial self-stimulation.

Chemical basis of reward and punishment

The next diagram shows how chemicals can be injected directly into the brain

In Stein's experiments an electrode was implanted into an animal's brain so that the tip of the electrode was located in the medial forebrain bundle. It is known that animals will press a lever at a very high rate for electrical stimulation from an electrode in this area.

A cannula (hollow tube) was implanted so that its tip lay in the ventricle of the brain. A liquid (Locke's solution which has a similar composition to naturally occurring cerebrospinal fluid) was passed down this cannula. Some animals had neurotransmitters (NA and 5-HT) dissolved in this liquid.

This technique allowed Stein to vary the concentration of neurotransmitter in the animal's brain and observe the effect of this on the rate of bar-pressing for electrical stimulation of the brain.

Stein hypothesized that

Effects of NA & 5-HT on reward systems
Effects of NA and 5-HT on ICSSThis diagram confirms Stein's suggestion that NA is involved in rewarding brain stimulation, whereas 5-HT is involved in punishment.

Compared to the baseline response rate:

The diagram shows facilitation of lateral hypothalamic self stimulation by norepinephrine (NE) and suppression by serotonin (5-HT).

Locke's solution is a placebo solution - it has the same composition as cerebrospinal fluid. Intraventricular injections were made 16 minutes after the start of the test.

Data are expressed as a per cent of the self-stimulation rate in the second 8 minute test period (8-16) min after start of test. The curves were obtained by averaging per cent scores for the group of 11 rats.

Stein's theory of reward & punishment

This diagram represents the hypothetical relationship between reward and punishment mechanisms in the brain and behaviour.

Signals of positive reinforcement release behaviour from periventricular system suppression by the following sequence of events:

(1) activation of norepinephrine (NE) containing cells in lower brain stem by stimuli previously associated with reward ( or avoidance of punishment) causes release of NE into amygdala and other forebrain suppresser areas via the medial forebrain bundle

(2) Inhibitory action of NE suppresses activity of the forebrain suppresser areas, thus reducing its cholinergically (Ach) mediated excitation of medial thalamus and hypothalamus

(3) Decreased cholinergic (Ach) transmission at synapses in medial thalamus and hypothalamus lessens the activity of the periventricular system, thereby reducing its inhibitory influence over motor nuclei of the brain stem.

Signals of failure or punishment increase behavioural suppression by the release of serotonin (5-HT), which either excites suppresser cells in the forebrain suppresser areas or disinhibits them by antagonizing the inhibitory action of NE.

From Stein et al, 1973 In Garatinni et al (Eds) The Benzodiazepines

Serotonin and anxiety

 Stein argued that this theory of the role of neurotransmitters in reward and punishment could be used to explain how benzodiazepines reduced and anxiety and explain why we feel anxious in certain circumstances.
According to Stein's theory:
  • anxiety is the result of increased serotonergic activity within the punishment system
  • benzodiazepine drugs reduce anxiety by reducing this increased serotonergic activity


The first strand of evidence which supports this theory comes from the observation that the anti-anxiety effect of benzodiazepine drugs takes time to develop. Anxiety does not disappear as soon as the patient starts taking the medicine, it develops over several days of treatment. Thus the reduction in anxiety is said to be a chronic effect of the drug.

In contrast many people report that they feel sedated shortly after beginning benzodiazepine treatment, but this disappears if the treatment is continued. Sedation is thought to be an acute effect of benzodiazepines.

There are short-term and long-term effects of benzodiazepines on neurotransmitter metabolism that parallel acute and chronic behavioural changes

This table summarizes the evidence which led Stein to suggest that

Stein argued that the chronic effect of BZPs - anxiety reduction - was due to reduction in serotonin turnover and that the recovery of NA turnover was responsible for the development of tolerance to sedative the side effects of BZPs.

There is evidence that the anti-conflict effects of BZPs in the Geller-Seifter paradigm is a chronic effect of BZP treatment.

Separation of chronic anti-anxiety & acute depressant effects of oxazepam in Geller-Seifter paradigm
This figure shows the increase of anti-anxiety activity and tolerance of depressant activity after repeated administration of oxazepam (20 mg/kg ip immediately before each daily test).

The cumulative record shows responding on the control (saline) day immediately preceding 22 days of daily oxazepam injections. The pen on the cumulative recorder resets every 3 minutes.

The numbers above the cumulative recordings represent the number of punished responses emitted by the rat.

Note the virtually complete suppression of unpunished responses after the first few doses of oxazepam (the sedative effect of the drug); the recovery of unpunished responding from day 3 onwards (indicating tolerance to the depressant effect of oxazepam) , and the increase in punished responding throughout the drug series (the anti-anxiety effect which does not exhibit tolerance with repeated injections).

Redrawn from Margules & Stein, 1968

Intraventricular 5-HT antagonizes anxiolytic effect of oxazepam
According to Stein the increase in punished responding produced by an anti-anxiety drug is due to the drug reducing serotonin activity. Stein tested this by attempting to reverse the effect of a BZP by injecting 5-HT directly into the brain.

The diagram shows the cumulative record of one rat performing on Geller-Seifter paradigm. Note that

This result is consistent with Stein's theory.

Action of drugs on serotonin
drugs and 5-HT As you know there are many ways of manipulating neurotransmitter activity within the brain. One way is to block synthesis of the neurotransmitter.

For example:

We can use these drugs to test Stein's theory.

According to Stein both drugs should have anxiolytic properties - increase responding on the CRF (punished) component on the Geller-Seifter conflict paradigm.

Anticonflict effect of chlordiazepoxide & methysergide

Sepinwall & Cook examined the effects of serotinergic drugs on conflict behaviour.

Their results are consistent with Stein's theory: Chlordiazepoxide (a BZP) and methysergide alone - and in combination - increased the number of punished responses made by rats on the Geller-Seifter paradigm

This slide shows the by now familiar effect of an anti-anxiety drug ( chlordiazepoxide) on punished responding in the Geller-Seifter paradigm.

Note that unpunished responding is unaffected by chlordiazepoxide indicating a specific effect of the drug on 'conflict' behaviour.

Methysergide (which blocks postsynaptic 5-HT receptors) on its own increases punished responding but it also increases non-punished behaviour (suggesting a non-specific increase in responding). Combining methysergide with chlordiazepoxide produces an even greater increase in punished responding. Data redrawn from Cook & Sepinwall, 1975.

Effect of PCPA & PCPA+5-HTP on punished responding
PCPA and PCPA+5-HTP effect on punished responding

This diagram shows the effects of PCPA and 5-HTP on conflict behaviour.

PCPA inhibits the enzyme tryptophan hydroxylase and thereby blocks 5-HT synthesis

5-HTP (5-hydroxytryptophan) is a serotonin precursor and was given after PCPA to restore the level of 5-HT in the brain.

As predicted by Stein's theory

The observation that responding declined after injection of 5-HTP suggests that the effect of PCPA on punished behaviour involved the serotoninergic system.

5-HTP fails to antagonize effect of diazepam on punishment
5-HTP and diazepamHowever this is not the end of the story. Kilts et al (1982) have challenged Stein's theory.

They found that a range of 5-HT antagonists at various doses did not affect punished responding in the predicted fashion.

One experiment (see Fig 3 in Kilts et al) produced results that were opposite to those predicted by the Stein theory.

The diagram shows the effects of 5-HTP, alone or in combination with a peripheral dopa decarboxylasse inhibitor (MK-486 increases the amount of 5-HTP getting to the brain) on the anti-conflict effect of a submaximal dose of diazepam (1.8 mg/kg).

The most significant finding is shown on the right hand side of the diagram.

Rats treated with 5-HTP+MK-486+diazepam made many more punished responses than animals treated with diazepam alone. In other words injecting drugs (MK-486+5-HTP) to increase serotonin activity in the brain did not antagonize the effect of diazepam, on the contrary boosting serotonin activity made the diazepam more effective than it would have been on its own.

This is exactly opposite to the result predicted by Stein's theory and I strongly recommend you read the Kilts et al(1982) paper .

In the 1980s attention switched from 5-HT to the role of another neurotransmitter GABA in the search for the mechanism underlying the anxiolytic effects of BZPs

Diazepam enhances effect of GABA
GABA is an inhibitory neurotransmitter. BZPs enhance inhibition at GABA synapses.

Electrical tracings compare the response of a mouse spinal cord neurone to GABA alone and in the presence of diazepam. Diazepam increases the effect of GABA which is an inhibitory transmitter that decreases the nerve membrane's excitability.

Benzodiazepines compete with GABA modulin to facilitate GABA receptor binding

BZPs compete with GABA-modulin for the GABA receptor site. The GABA receptor complex is illustrated and explained in Lickey & Gordon, page 288 and in the article by Ashton.

This diagram shows an hypothetical GABA receptor complex.

Attached to the high-affinity GABA receptor is a satellite receptor for GABA modulin .

When GABA modulin binds to the satellite receptor site the receptor closes thereby blocking GABA access to its receptor site .

The lower part of the diagram shows how benzodiazepines compete with GABA modulin for the satellite site but they do not activate (close) it thus leaving GABA free access to its receptor site.

GABA receptor binding and clinical potency

But the correlation between a BZPs ability to bind to the GABA receptor and its clinical potency is relatively low

Drug affinities for benzodiazepine receptors correlates with the drug's anti-anxiety potency but only weakly suggesting that this mechanism is not a complete account of how the drugs reduce anxiety.

Receptor binding correlates weakly with potency in animal model (Geller-Seifter paradigm)

Similarly the correlation between a BZPs ability to bind to the GABA receptor and its potency in the Geller-Seifter paradigm is relatively low

Drug affinities for benzodiazepine receptors correlates weakly (r=0.661) with the drug's anti-anxiety potency as measured by the median effective dose (MED) in animal tests. Further evidence that receptor binding is not a complete answer to how anxiolytics effect anxiety. Recall that a drug's potency on the Geller-Seifter paradigm is a very good predictor of clinical potency

Unanswered questions

beta-CCM has anxiogenic properties and antagonizes the effects of BZPs in a conflict test (see de Carvahlo et al,1983).

Nociceptin is found in the brain. It is structurally very like the opium category of drugs, but it does not act through the same pathways. It may be a naturally occurring anti-anxiety drug. A team led by Francois Jenck from Roche CNS Research in Basel, Switzerland, reported in late 1997 that nociceptin produces a general reduction in anxiety in animal models of anxiety.

For example, rats and mice are normally very wary of bright lights and open spaces. When given small doses of nociceptin, they showed a reduced tendency to hide in dark corners. The animals were also tested in a 'conflict' situation. Animals treated with nociceptin were much more likely to ignore the stress of the shock and go for the food.

You can read more about this research in this Nature article Brain: Anxiety about brain chemistry by Helen Phillips

Links between GABA and other neurotransmitter systems
Recent research has returned to examining the role of 5-HT in anxiety - this resurrection of interest in serotonin was because we now know that there is more than one type of serotonin receptor.

We know that benzodiazepines potentiate the GABA system, it may turn out that GABA exerts inhibitory effects on other neurotransmitter systems which may account for the behavioural effects of anxiolytics

Chemicals that antagonize the effects of BZPs are interesting because alcohol and anxiolytics are similar in many ways (see Lickey & Gordon) and it would be useful to have a drug that rapidly reversed the effects of alcohol for example to use in hospital emergency departments.

The references give a flavour of current research on the BZP antagonist flumazenial and its ability to reverse the effects of alcohol.

'Second generation' anxiolytics & the re-emergence of 5HT in anxiety

Chlorazepate compared with buspirone The diagram shows the effects of clorazepate, a benzodiazepine, and buspirone on anxiety symptoms during drug treatment and drug withdrawal in anxiety patients.

Patients received drug treatment for 24 weeks and were then switched to placebo for four weeks of withdrawal at the end of drug treatment. Note that patients respond more slowly on buspirone compared to the benzodiazepine. But in withdrawal there is a 'rebound' of anxiety in the benzodiazepine group, but not in patients treated with buspirone. Therefore the withdrawal symptoms are evoked by benzodiazepine withdrawal and are not due to the return of the original illness. Figure redrawn from Lickey & Gordon, Medicine and Mental Illness, 1991.

Clearly buspirone is not a magic bullet for anxiety, it has the advantage that it does not produce withdrawal symptoms, but its slow onset is a disadvantage because the patient may give up on taking the medicine if there is no early signs of symptom reduction.

Buspirone is called a ' second generation ' anxiolytic because - unlike the older benzodiazepine drugs - it binds selectively to a subclass of serotonin receptors - so called 5HT 1A receptors. Consequently there there is a need to revisit the role of serotonin in anxiety, a theory put forward by Larry Stein in the 1970s which was explored at the heart of this lecture

Point to ponder
Can you think of an animal model of anxiety? Why would it be better than the Geller-Seifter conflict paradigm? How would you experimentally verify this claim?

Online resources


Glossary & Abbreviations:

Supplementary information

  • American Psychiatric Association Benzodiazepine Task Force Report On Use, Dependence, Toxicity and Abuse . 1989
    In the USA, 11-15 % of the general population (not just anxiety patients) use bz's on an as needed (prn) basis.
    • 80% of those people use them every day for less than 4 months,
    • 67% - every day for less than 1 month. This includes people hospitalized for medical reasons. These bz's are prescribed mostly by family practitioners and internists for medical/surgical use.
    • 15% of all users take bz's regularly for a year or more. Older, medically ill patients are the most common users; they tend not to increase dosage and both the patients and their physicians agree the bz's are helpful.
  • Review of Behavioral Effects of Benzodiazepines With an Appendix on Drawing Scientific Conclusions from the FDA's Spontaneous Reporting System (MedWatch), By Peter R. Breggin, M.D.
  • Excerpts from Toxic Psychiatry , Chapter 11 by Peter R. Breggin, M.D.
Copyright Dr. C.A.P. Kenyon 1994-2006