Systematic desensitisation

1. Introduction

Version followed: Joseph Wolpe's relaxation-based imaginal systematic desensitisation as set out in The Practice of Behavior Therapy (4th ed., 1990), with the reciprocal-inhibition rationale from Psychotherapy by Reciprocal Inhibition (1958). Named variants (in vivo, self-control, emotive imagery, group, automated) are marked as such. Caution: the 1990 book text is not drawn on here. Its procedural parameters come from a detailed university teaching summary that cites it by page (Athabasca University). Points not drawn from a cited source are tagged [uncertain].

Systematic desensitisation (SD) treats conditioned fears. The client is taught a response that is physiologically incompatible with anxiety, usually deep muscle relaxation. Then, while in that state, the client is shown the feared stimuli in small graded steps, usually in imagination. Wolpe's reciprocal inhibition principle says that if a response antagonistic to anxiety happens while anxiety-evoking stimuli are present, and suppresses the anxiety, the bond between those stimuli and anxiety is weakened. Each weak, fully inhibited dose of anxiety reduces the next item's anxiety, so the client can climb a hierarchy without being overwhelmed. "Reciprocal inhibition therapy" is Wolpe's wider label. It also covers other anxiety-inhibiting responses: assertive, sexual, respiratory and feeding responses. These are listed in Section 4. Later research found that graded imaginal exposure works about as well with or without relaxation. Many clinicians now therefore explain SD as graded exposure. This guide teaches Wolpe's procedure as written. [source: Wolpe 1958; Wolpe 1990 via Athabasca; Carona et al. 2023; Wikipedia]

2. Framework

SD has three components that partly overlap in time. Behaviour analysis comes first. [source: Athabasca concept page; Wikipedia; Carona et al. 2023]

Phase 0: Behaviour analysis (assessment), sessions 1 to 2 or more. Goal: identify exactly which stimuli trigger anxiety and through what learning history. Separate fears that are classically conditioned (good fit for SD) from fears that rest on misinformation. Misinformation fears need correction first. Check for other targets such as unassertiveness or interpersonal anxiety. Tools: detailed history, stimulus-by-stimulus questioning, Wolpe & Lang's Fear Survey Schedule, and the 0–100 SUD scale. Move on when: there is a working list of anxiety themes and the client accepts the learning rationale. [source: Athabasca; Wolpe Wikipedia ("behavior analysis")] [uncertain: the conditioned vs. misinformation distinction is from my recollection of Wolpe 1990]

Phase 1: Relaxation training, about 6 lessons. Goal: the client can reach deep relaxation quickly from a brief cue. Format: each lesson takes about the first third of a session. Home practice is twice a day, 10–15 min. This is an abbreviated version of Jacobson's progressive relaxation, which used 50–200 sessions. Hierarchy work runs alongside in the rest of each session. Move on when: the client reliably relaxes to near-zero SUD within a few minutes. [source: Athabasca; Scherer & Pass]

Phase 2: Hierarchy construction, done alongside Phase 1. Goal: one or more themed hierarchies, usually 10–20 items (some texts say 10–30), ranked by the client's own SUD ratings and spaced about 5–10 SUD apart. Move on when: each hierarchy has a low starting item that the client expects to cause only mild disturbance, and no large gaps between items. [source: Athabasca, citing Wolpe 1990 p. 164; Miltenberger 1997; Martin & Pear 1998]

Phase 3: Desensitisation proper. Goal: every item, up to the top one, can be imagined at SUD 0. Format: the desensitisation part is about 15–30 min of a session. Early sessions use 8–10 scene presentations and later ones 30–50. Up to four hierarchies can be worked in one session. A scene is usually given 3–4 times before it reaches 0, sometimes 10 or more. Move between items when: the current scene produces no disturbance (SUD 0) on repeat presentations. Frequency and dose: spacing matters little, but at least weekly is recommended. Total sessions range from 1 to about 100 in Wolpe's reports. For a simple phobia, practice summaries suggest about 4–12. [source: Athabasca; MLSU notes; Wikipedia] [uncertain: the 4–12 range comes from secondary teaching notes, not Wolpe]

Phase 4: Transfer and generalisation. Goal: reduced fear in the real situation. Format: check real-life fear. Where imaginal gains do not transfer, add in vivo desensitisation. Wolpe estimated that about 15–20% of patients need it. End when: the agreed goals are met, which may not mean every trace of fear is gone. [source: Athabasca variations page, citing Wolpe 1982; Miltenberger 1997; MLSU notes]

Setting: individual outpatient sessions, delivered by a behaviour or CBT therapist. Division 12 lists SD as an accepted but slower alternative to in vivo exposure. It suits clients who will not yet accept direct exposure, and fears that are impractical to recreate (thunderstorms, flying, rare events). [source: Div 12]

3. Guiding principles

Anxiety is a learned habit, not a character flaw. Present the fear as conditioned and therefore unlearnable. Keep a non-judgemental, matter-of-fact stance. [source: Wolpe 1958 rationale; Carona et al. 2023] [synthesis: stance wording]
Reciprocal inhibition: the competing response must be stronger than the anxiety. Only present a scene the relaxation can overpower. If the client is disturbed, the step was too big. [source: Wolpe 1958, p. 139 via Athabasca]
Small doses, fully neutralised. Each presentation should end with anxiety extinguished, not endured. That is why scenes are brief (about 5–7 s) and stopped at any sign of disturbance. [source: Athabasca; Scherer & Pass]
The client's ratings decide the order. Rank items by the client's own SUDs, even when the order looks illogical to you. [source: Athabasca]
SUDs are within-person yardsticks. Use them to track one client over time. Do not compare clients. [source: Athabasca, citing Wolpe 1990]
Precision of stimulus. Desensitise the stimulus that actually drives the fear, not the one it looks like at first. Find that stimulus by careful behaviour analysis. [source: Wolpe Wikipedia (misleading hierarchies as a known failure); Athabasca]
Never advance past disturbance. Move up only when the current item reaches 0. When an item gets stuck, go back down or insert steps. [source: Athabasca]
Relaxation is a means, not the essence. Wolpe did not consider relaxation indispensable. Any response that inhibits anxiety can serve, such as assertion, sexual arousal, eating, pleasant emotion or humour. Pick one that fits the client. [source: Athabasca; Wolpe Wikipedia]
End each session on success. Close on a scene the client managed calmly. Start the next session at the unfinished item. [source: Athabasca]
Transfer is the real test. Imaginal gains must be checked in vivo and supplemented if needed. [source: Athabasca, citing Wolpe 1982; Div 12]
Client keeps control. The client signals and can always stop a scene. This builds safety and consent. [source: Athabasca (finger signal)] [synthesis: framing as consent]

4. Intervention inventory

This section has three parts. A covers the core Wolpe procedure (items 1–11). B covers published variants of SD (items 12–18). C covers the other reciprocal-inhibition procedures Wolpe grouped with SD (items 19–23). All example language is my own illustrative wording; none of it is Wolpe's text. [source: Athabasca concept/variations pages; Wikipedia; Wolpe 1958, 1990]

A. Core procedure

1. Behaviour analysis and rationale

Purpose / when: first 1–2 sessions. Find the real eliciting stimuli and how the fear was acquired, and spot other targets. Then give a learning-based rationale that earns consent. [source: Wolpe Wikipedia ("behavior analysis"); Athabasca]
Steps: (1) Take the presenting complaint and ask for the most recent episodes in sequence: where, who, what was seen or heard, the first bodily sign. (2) For each episode, change one variable at a time (distance, time to the event, number of people, who is present) and ask how the fear rises or falls. This yields the dimensions for the hierarchy. (3) Take the learning history: the first occurrence and any traumatic or repeated pairings. (4) Ask whether the fear rests on a factual belief that is wrong (e.g., that panic causes heart attacks). Correct that with information before desensitising [uncertain: recollection of Wolpe 1990]. (5) Screen for related areas: assertiveness, sexual function, general tension. (6) Give the rationale: the fear was learned by association, it can be unlearned by pairing the same cues with calm in steps too small to overwhelm, and the client controls the pace.
Example language (illustrative): "Your body learned to sound the alarm around this cue. We'll teach it a different response, a few small steps at a time. You'll never be asked to go further than you can handle calmly."
Watch for: fears that are really about something else ("fear of crowds" turning out to be fear of being stared at). A good analysis gives dimensions you can turn up or down in small steps.

2. SUD scale training

Purpose / when: set up a shared measure before relaxation and hierarchy work. Use it in every scene trial. [source: Athabasca, citing Wolpe 1990]
Steps: (1) Define the anchors: 0 is complete calm and 100 is the worst anxiety the client has ever felt or can imagine. (2) Ask for a few reference points from the client's life (a calm Sunday, a mildly tense meeting, the worst episode) and note their numbers. (3) Ask for the client's SUD right now. (4) Explain that the numbers only need to make sense to the client, and that rating gets easier with practice.
Example language (illustrative): "On a scale where 0 is totally at ease and 100 is the most upset you've ever been, where are you sitting right now?"
Watch for: ratings that cluster (everything 80–100). If so, re-anchor with more personal reference points. Wolpe reported that SUDs correlate with physiological measures, which supports using them. [source: Athabasca]

3. Fear Survey Schedule and questionnaires

Purpose / when: during assessment, to find fears the client didn't mention. [source: Athabasca, citing Wolpe 1990 Appendix C]
Steps: (1) Give the Fear Survey Schedule (Wolpe & Lang). It is a self-rated list of common fear objects and situations. Do not reproduce it; use the published form. (2) Go through high-rated items with the client to see whether they belong to the presenting theme or form separate hierarchies. (3) Wolpe also used general neuroticism and self-sufficiency questionnaires [uncertain: Willoughby and Bernreuter scales recalled].
Example language (illustrative): "You marked several situations involving being watched. Let's see whether they're part of the same fear."
Watch for: high scores across many items. That points to general tension, so consider strengthening relaxation and screening for generalised anxiety.

4. Abbreviated progressive relaxation training

Purpose / when: sessions 1–6 or so, about the first third of each session. It builds the competing response. [source: Athabasca, citing Wolpe 1990; Scherer & Pass]
Steps: (1) Seat the client in a reclining chair, eyes closed if comfortable. (2) Work through muscle groups in a fixed order. Wolpe's groups are: arms; face (forehead, eyes); jaw and tongue; neck and shoulders; back, abdomen and chest (including calm breathing); legs and feet. Cover about one or two regions per lesson and keep reviewing the earlier ones. (3) For each group, have the client tense moderately for several seconds, notice the tension, then release and attend to the contrast as the muscles let go further, past the point of ordinary rest. (4) Over the lessons, drop the tensing so that release alone, then a cue word, brings on relaxation. (5) End each lesson with a SUD rating and a slow return to alertness.
Example language (illustrative): "Tighten the right fist… notice the strain in the forearm… now let it go. Keep letting go, further than you would normally."
Watch for: muscle cramps (lighten the tension), self-monitoring ("am I doing it right?"), and relaxation-induced anxiety (see Section 5). It is working when SUD falls to about 0–10 within a few minutes of the cue.

5. Home relaxation practice

Purpose / when: throughout Phase 1 and after. [source: Athabasca]
Steps: (1) Assign two practices a day of 10–15 min each, covering the groups learned so far. (2) Ask for a simple log: time, start SUD, end SUD, any problems. (3) Review the log at the start of the next session and fix problems before moving on.
Example language (illustrative): "Twice a day, run through the groups we did today. Note your starting and ending number."
Watch for: practice done only when the client is already calm, or skipped. Problem-solve when and where to practise.

6. Differential and cue-based relaxation in daily life

Purpose / when: once basic relaxation is learned. Lowers background tension so that hierarchy scenes start from a calm base. [uncertain: Jacobson's differential relaxation is standard, but how much Wolpe used it is not established]
Steps: (1) Practise relaxing the muscles not needed for a task, e.g., while sitting, then talking, then walking. (2) Pair a word such as "calm" with the exhale at the end of each practice so the word becomes a cue. (3) Use the cue in mildly tense daily moments.
Example language (illustrative): "Keep your hand on the pen but let your shoulders and jaw go."
Watch for: using the cue to escape or avoid situations. Its purpose is to lower tension, not to avoid. [synthesis]

7. Hierarchy construction

Purpose / when: alongside relaxation training. It sets the graded steps. [source: Athabasca, citing Wolpe 1990 p. 164; Miltenberger 1997; Martin & Pear 1998]
Steps: (1) Brainstorm situations within one theme, including ones the client has never been in but expects to fear. (2) Write each on a card with concrete detail (who, where, what is happening). (3) The client rates each card in SUDs. (4) Order the cards by rating, keeping the client's order even when it seems illogical. (5) Check spacing. Aim for about 5–10 SUD between items and add items where the gap is larger. Wolpe warned that items too close together waste time. (6) Make sure the lowest item causes only slight disturbance, about 10–20. If not, invent an easier item. (7) Build a separate hierarchy for each theme; several can run in parallel. (8) Hierarchies often vary one dimension, such as distance, time before the event or number of people. Wolpe distinguished spatio-temporal hierarchies from thematic ones [uncertain].
Example language (illustrative): "If 'standing at the window of the third floor' is 40 and 'tenth floor' is 70, what would be a 55?"
Watch for: items that are vague, combine two fears, or are at the wrong level. These show up later as items that won't come down (see Section 5). The finished hierarchy usually has 10–30 items.

8. Imagery check with a neutral control scene

Purpose / when: at the start of desensitisation proper. Tests whether the client can form a clear image and drop it on instruction. [source: Athabasca]
Steps: (1) Relax the client. (2) Present a neutral scene unrelated to the fear, e.g., a familiar street corner on an ordinary day. (3) Ask the client to signal when it is clear, then to stop the scene. (4) Ask for the SUD (expect 0) and how vivid the image was. (5) If the image is weak or will not clear, see Section 5 (poor imagers).
Example language (illustrative): "Picture yourself at a familiar corner on an ordinary morning… raise your finger when you have it… now stop the scene and just relax."
Watch for: any disturbance to a neutral scene. That suggests generalised tension, or that relaxation itself is producing anxiety.

9. Scene presentation (desensitisation proper)

Purpose / when: the central intervention, used from the end of relaxation training to the top of each hierarchy. [source: Athabasca, citing Wolpe 1990; Athabasca example page; Scherer & Pass]
Steps: (1) Agree the signals beforehand. The client raises the left index finger about an inch when the scene is clear. The client can signal disturbance or stop at any time, and stops imagining immediately if anxiety becomes strong. (2) Relax the client with the abbreviated cue procedure and check the SUD (aim near 0). (3) Present the control scene first (item 8). (4) Describe the lowest hierarchy item briefly and concretely and ask the client to imagine being there. (5) When the finger rises, let the scene run about 5–7 seconds, then say "stop the scene". (6) Ask for the SUD the scene produced (a number or finger signal). (7) Renew relaxation for about 20–30 seconds; between scenes leave about 10–30 seconds. (8) Present the same scene again. Usually the SUD falls across 3–4 presentations. (9) When it gives 0, normally on two successive presentations [uncertain: the exact count of zero trials varies between accounts], move to the next item. (10) Early sessions use 8–10 presentations, later ones 30–50, and up to four hierarchies in one sitting. The whole sequence takes about 15–30 min. (11) End on a scene the client managed calmly. Record where you stopped. The next session starts with the unfinished item, or with the next item up if the last one was completed.
Example language (illustrative): "Keep relaxing… Now imagine you're standing in the lift doors, the door is open, and two people are inside… [finger] … hold it… stop the scene. How much did that disturb you? … Now let go of it; let your arms and shoulders relax again."
Watch for: any rise of the finger or any SUD above 0 on an item that should be easy. Do not push on through it. Watch the face, breathing and hands for tension the client does not report. Signs that it is working: the SUD per item falls across presentations, each new item starts lower than its rating on the card, and the client reports less avoidance in daily life.

10. Remedial manoeuvres for an item that won't fall

Purpose / when: an item still produces disturbance after several presentations, or the SUD rises. [source: Athabasca]
Steps: (1) Shorten the exposure: present the same scene for only 2–3 seconds. (2) Go down to the previous item and give it longer presentations (e.g., 20 seconds) to strengthen the calm. (3) Insert intermediate items between the last calm item and the stuck one by changing a single detail (more distance, fewer people, less time). (4) Deepen relaxation before the next presentation. (5) If none of these helps, re-examine the hierarchy (Section 5, "Misleading hierarchy").
Example language (illustrative): "Let's step back to the one you handled and hold it a bit longer… Now we'll make the next one a smaller step."
Watch for: repeated failure on the same item. That usually means the item is mis-specified, not that the client lacks effort.

11. Hypnotic induction as a relaxation aid

Purpose / when: historically, to deepen relaxation for clients who could not relax enough. [uncertain: Wolpe 1958 reports using hypnosis for many desensitisation patients; recalled]
Steps: use only if you are trained in clinical hypnosis. Induce, then present scenes as in item 9. Contemporary practice usually relies on relaxation training alone. [synthesis]
Example language: not supplied; follow your hypnosis training.
Watch for: dependence on the therapist's induction, which works against the client becoming self-sufficient.

B. Variants of systematic desensitisation

12. In vivo desensitisation

Purpose / when: when imagined scenes do not provoke the real emotion (Wolpe estimated about 15–20% of patients), when imaginal gains don't transfer, or when the client cannot form images. It also avoids the generalisation problem because the work happens in the real situation. [source: Athabasca variations page, citing Wolpe 1982 and Miltenberger 1997; Div 12]
Steps: (1) Turn the hierarchy into real tasks graded the same way. (2) Get informed consent covering the plan, the right to stop, confidentiality in public settings and any physical risks (see Section 5, safety). (3) Accompany the client at first; the therapist's presence is often itself calming. (4) The client approaches the step while relaxing, rates SUD, and retreats or pauses if the SUD climbs beyond mild. Then they return. (5) Repeat until the step is calm, then move up. (6) Gradually hand the steps to the client as homework.
Example language (illustrative): "We'll walk to the corner where you can see the bridge. Stop whenever you need, breathe out and let your shoulders drop, and tell me your number."
Watch for: safety behaviours, such as gripping or not looking, that keep anxiety down without real contact. Check that gains hold without you present. Some targets cannot be accompanied (e.g., intimacy with a partner) and must be assigned as structured homework. [source: Athabasca]

13. Contact desensitisation / participant modelling (guided mastery)

Purpose / when: in vivo work where the client needs modelling and physical guidance. It is mainly used for animal phobias, heights and driving. [source: Div 12 (participant modelling / guided mastery; Bandura, Blanchard & Ritter 1969; Williams et al. 1984, 1985)] [uncertain: the detailed steps below are not from a cited manual]
Steps: (1) The therapist models the step, e.g., touching the container and then the animal. (2) The client does it with physical support, such as the therapist's hand on theirs. (3) Fade the support. (4) Move up the hierarchy.
Example language (illustrative): "Watch me first… now put your hand on mine while I touch it… now try it on your own."
Watch for: the client depending on the therapist's presence. Fade it on purpose.

14. Self-control desensitisation (Goldfried 1971)

Purpose / when: when you want the client to learn a general coping skill rather than pure counterconditioning, e.g., for multiple or generalised anxieties. [source: SAGE CBT encyclopedia entry; Scherer & Pass]
Steps: (1) Give the rationale: relaxation is a coping skill you will learn to use whenever tension starts. (2) Train relaxation and slow breathing, plus attention to early bodily tension cues. (3) Build scenes together. They can come from different situations and need not be strictly graded. (4) The client imagines a scene for about 10–15 s. (5) If anxiety appears, the client keeps imagining and "relaxes away" the tension. The scene is ended only if anxiety becomes unmanageable. (6) Repeat until the scene is handled easily. (7) Assign homework applying the skill in real situations and self-presenting scenes.
Example language (illustrative): "When you notice the tightening, stay in the scene and let the tension drain out as you breathe out."
Watch for: this reverses Wolpe's rule of stopping at the first anxiety. Do not mix the two instructions within one hierarchy.

15. Emotive imagery for children (Lazarus & Abramovitz 1962)

Purpose / when: children who can't do formal relaxation. Positive excitement from imagined stories is used as the competing response. [source: Lazarus & Abramovitz 1962 abstract (reciprocal-inhibition basis)] [uncertain: the procedural steps below are from the wider literature, not the abstract]
Steps: (1) Build a graded hierarchy with the child and parent. (2) Find the child's favourite hero or story character and what excites them about it. (3) Tell a story in which the child is with the hero on an adventure that leads into the lowest feared situation, keeping excitement and pride high. (4) If the child shows fear, switch back to the exciting story thread. (5) Move gradually up the hierarchy within later stories.
Example language (illustrative): "You and Captain Star are on a secret mission. The signal says the treasure is in the dark hallway… Captain Star says you're the bravest helper he has."
Watch for: the child disengaging from the story or becoming afraid. Shorten the step and boost the hero thread. Check with the parent for real-life change.

16. Group desensitisation

Purpose / when: several clients with the same fear, e.g., public speaking, test anxiety, spiders. [source: Athabasca variations page (Paul & Shannon 1966; Ihli & Garlington 1969; Robinson & Suinn 1969); Div 12 (Lazarus 1961)]
Steps: (1) Teach relaxation as a group. (2) Build a shared hierarchy from pooled items and ratings. (3) Present scenes to the whole group. Any member signals disturbance, and that scene is repeated or the group drops back. The pace is therefore set by the slowest member [uncertain: the pacing rule is from the wider literature, not a cited source]. (4) Members practise individually at home.
Example language (illustrative): "If anyone feels any disturbance, raise a finger and we'll go over that scene again."
Watch for: members who don't signal so as not to hold the group back. Check SUDs privately on paper.

17. Automated or self-administered desensitisation

Purpose / when: to save therapist time, or for between-session practice. [source: Athabasca variations page (Migler & Wolpe 1967; Chandler et al. 1986; Martin & Pear 1998 ch. 25)]
Steps: (1) Record relaxation instructions and scene descriptions in hierarchy order, with pauses. (2) The client controls playback, repeating a scene until calm before moving on. (3) Review logs in session.
Example language (illustrative): "Play scene 4 only after scene 3 has been at zero twice."
Watch for: skipping ahead. Check SUD logs, and keep in-session checks.

18. Alternative competing responses

Purpose / when: relaxation is unsuitable or fails. Wolpe did not treat relaxation as essential. [source: Athabasca; Wolpe Wikipedia]
Steps: choose a response the client can produce strongly in the presence of the low-level stimulus. Options are assertive speech (item 19), sexual arousal (item 20), eating (historically with children, after Jones), pleasant or humorous imagery, physical activity, or other relaxation methods such as autogenic training, meditation, yoga or biofeedback. Pair it with graded items exactly as in item 9.
Example language (illustrative): "As you picture the meeting, bring back the feeling of that holiday morning we talked about."
Watch for: whether the response actually inhibits the anxiety (SUD drops). A response that is too weak simply exposes the client to unbuffered anxiety.

C. Other reciprocal-inhibition procedures (Wolpe 1958, 1990)

The following are part of Wolpe's broader reciprocal inhibition therapy but are not SD itself. Items 21 and 22 are listed for completeness and are not recommended for routine use today. [uncertain: the Wolpe 1958 and 1990 chapters' contents are recalled and partly confirmed by the Wolpe Wikipedia page for assertion and relaxation]

19. Assertive training (behaviour rehearsal)

Purpose / when: anxiety in interpersonal situations, especially where the client cannot express legitimate wishes, refusal or irritation. Expressing these is thought to inhibit the anxiety. [source: Wolpe Wikipedia (assertiveness as an anxiety-inhibiting response, useful for social anxiety)]
Steps: (1) Identify specific situations where the client stays silent, over-complies or explodes. (2) Explain the rationale: the anxiety keeps the response suppressed, and expressing the appropriate feeling reduces the anxiety over time. (3) Role-play: the therapist plays the other person and the client responds. (4) Give feedback on content, voice, eye contact and posture. Model a better version, then repeat. (5) Grade the real-life assignments from low-risk situations (e.g., a shop) to high-risk ones (e.g., family). (6) Review the outcomes and adjust.
Example language (illustrative): "I'll be your colleague who keeps leaving work on your desk. Tell me, in one or two sentences, that you won't take it on."
Watch for: aggression mistaken for assertion; assignments with real danger (e.g., abusive partners, where you should not assign confrontation, but plan for safety [synthesis]); and cultural norms about directness. It is working when the client's SUD in role-play falls and they report acting assertively in real life.

20. Sexual responses for sexual anxiety (graded in vivo)

Purpose / when: sexual dysfunction maintained by performance or intimacy anxiety, with a cooperative partner. Sexual arousal serves as the competing response. [uncertain: Wolpe 1958 describes this]
Steps: (1) Assess the problem and rule out medical causes with the client's physician [synthesis]. (2) Explain the rationale to both partners. (3) Agree that the couple will go only as far in physical intimacy as the client can while pleasant feelings stay clearly stronger than anxiety. Stop or step back if anxiety dominates. (4) Increase intimacy gradually over successive occasions. (5) Review at each session. This approach anticipates later sensate-focus programmes. [synthesis: link to sensate focus]
Example language (illustrative): "The only rule is to stop at the point where enjoyment stops being the main feeling. Next time, you can go a little further."
Watch for: pressure from the partner, and relationship conflict that needs couples work first. [synthesis]

21. Respiratory responses (historical: carbon-dioxide/oxygen inhalation)

Purpose / when: historically used by Wolpe for pervasive, free-floating anxiety. [source: Carona et al. 2023 (respiratory responses for free-floating anxiety)]
Steps: none given. This was a medically supervised procedure, has no place in routine psychotherapy, and I do not describe it operationally. For pervasive anxiety, current practice uses CBT for generalised anxiety or applied relaxation instead. [synthesis]
Watch for: n/a.

22. Anxiety-relief conditioning (historical)

Purpose / when: historically, a cue word was paired with the ending of an unpleasant electrical stimulus, so that the word would later bring relief. [uncertain: recalled from Wolpe 1958/1990]
Steps: none given. Aversive stimulation is not acceptable in current routine practice and has little evidential support. Use cue-controlled relaxation (item 6) instead. [synthesis]
Watch for: n/a.

23. Adjacent techniques in Wolpe's textbook (not SD)

Purpose / when: The Practice of Behavior Therapy also covers techniques that are not reciprocal inhibition, such as thought stopping and flooding. They are outside this guide. Thought stopping in particular is now generally discouraged because of research on thought suppression. [uncertain: inclusion recalled; evaluation is synthesis]
Steps / language / watch for: see the relevant exposure or cognitive guides.

5. Handling problems

Tags: [source] means the guidance comes from the sources; [synthesis] means my own clinical synthesis.

Safety and consent before any exposure (imaginal or in vivo). (1) Get informed consent covering what will happen, the expected temporary discomfort, the client's control over pace through the stop signal, and alternatives such as in vivo exposure. Division 12 notes that in vivo exposure is generally the first-line option if the client will accept it. (2) Contraindications and cautions: active psychosis; current intoxication; acute suicidality (assess using standard risk assessment and stabilise first); unstable cardiac or respiratory illness, pregnancy or seizure disorder before strenuous in vivo tasks (consult the client's physician); severe dissociation; and ongoing real danger, which should be dealt with as safety, not as fear. (3) Blood-injection-injury phobia: clients may faint from a drop in blood pressure. Use Öst's applied tension rather than relaxation, because relaxation may make the drop worse. (4) Debrief after each session and check the SUD before the client leaves. [source: Div 12 (first-line status, applied tension); synthesis for the rest]

Client can't relax / relaxation-induced anxiety. Lengthen training. Try a different method (breathing, autogenic training, meditation, biofeedback). Do the training eyes open. Shorten the relaxation segments. Or switch to another competing response (item 18) or to plain graded exposure, which works about as well without relaxation. [source: Athabasca; Carona et al. 2023; Scherer & Pass] Referral for medication to reduce baseline anxiety is a historical option Wolpe mentioned [uncertain]. Do not dose; liaise with a prescriber.

Poor imager (weak, fleeting or emotionless scenes). (1) Do imagery practice with neutral and pleasant scenes, adding sensory detail: sounds, smells, body position. (2) Have the client describe the scene aloud in the first person and present tense. (3) Use photographs, video or recorded sounds as stimuli. (4) Move to in vivo (item 12) or VR exposure. Wolpe originally used real stimuli, and imagined scenes do not work for everyone. [source: Wikipedia (aphantasia, VR); Athabasca variations; Wolpe Wikipedia (imagery difficulties as a known problem)] [synthesis: steps 1–2]

Misleading hierarchy (items don't move, or gains don't generalise). Go back to behaviour analysis. The real stimulus may differ from the surface one, for example fear of judgement rather than of crowds, or fear of losing control rather than of heights. Rebuild the hierarchy around the true dimension. [source: Wolpe Wikipedia (misleading hierarchies); synthesis for the example]

SUD spikes on an item. Stop the scene at once and relax. Go back to the last calm item and end there. Next time, insert intermediate steps (item 10). Do not end the session on a high SUD. [source: Athabasca]

Client avoids within the image (imagines watching from a distance, switches the scene, "blurs" it). Ask what exactly they pictured. Re-specify the scene in the first person, with concrete details. Acknowledge that avoiding is understandable, then shrink the step instead of letting the avoidance continue. [synthesis]

Overwhelm / panic in vivo. Pause in place if possible and let the anxiety settle with breathing or relaxation. Step back one level, then return. Avoid an abrupt escape if possible, but never force the client. Plan the next step smaller. [synthesis]

Non-adherence (no home practice, skipped in vivo tasks). Ask what got in the way, then shrink and schedule the task: a specific time, place and cue. Use a log, review it at the start of every session, and praise partial completion. Check whether the task was too high on the hierarchy. [synthesis]

Stuck: progress plateaus over sessions. Check the following: in-session SUDs are actually reaching 0; between-session avoidance or safety behaviours are undoing the gains; there are untreated co-occurring problems such as depression, unassertiveness or substance use; the client's environment is reinforcing the fear (e.g., family accommodation). Consider switching to in vivo exposure, which Division 12 rates more highly and finds faster. [source: Div 12] [synthesis for checklist]

Resistance or doubt about the rationale ("imagining won't change anything"). Normalise the doubt. Show the data from their own SUD ratings across presentations. Offer a short trial with explicit review points. If the client prefers it, offer direct in vivo work. [synthesis]

Rupture (client feels pushed, misunderstood or embarrassed). Name it, take responsibility for your part (e.g., "I moved us too fast"), and give control back to the client over pacing and item choice. Repair before continuing the hierarchy. [synthesis]

Multiple fears / generalised anxiety. Build parallel hierarchies; up to four can be worked in one session. Prioritise by impairment. Consider self-control desensitisation (item 14) for diffuse anxiety. [source: Athabasca; SAGE entry]

Children. Use emotive imagery (item 15) or in vivo steps with a parent present. Keep sessions short and use concrete rewards. [source: Lazarus & Abramovitz 1962] [synthesis for session structure]

Trauma-related fears. SD is not a substitute for trauma-focused treatments recommended in current guidelines (e.g., prolonged exposure, CPT, EMDR, trauma-focused CBT). Use SD only for discrete residual fears, after appropriate assessment. [synthesis]

Sources

Sources:

Further reading (not drawn on directly): books and papers cited

Wolpe, J. (1958). Psychotherapy by Reciprocal Inhibition. Stanford University Press.
Wolpe, J. (1961). The systematic desensitization treatment of neuroses. Journal of Nervous and Mental Disease, 132, 189–203.
Wolpe, J., & Lazarus, A. A. (1966). Behavior Therapy Techniques. Pergamon.
Wolpe, J. (1990). The Practice of Behavior Therapy (4th ed.). Pergamon. (Earlier eds. 1969, 1973, 1982.)
Jacobson, E. (1938). Progressive Relaxation. University of Chicago Press.
Goldfried, M. R. (1971). Systematic desensitization as training in self-control. Journal of Consulting and Clinical Psychology, 37, 228–234.
Migler, B., & Wolpe, J. (1967). Automated self-desensitization. Behaviour Research and Therapy, 5, 133–135.
Paul, G. L., & Shannon, D. T. (1966). Treatment of anxiety through systematic desensitization in therapy groups. Journal of Abnormal Psychology, 71, 124–135.
Öst, L.-G., & Sterner, U. (1987). Applied tension. Behaviour Research and Therapy, 25, 25–29.
Miltenberger, R. (1997). Behavior Modification: Principles and Procedures. Brooks/Cole.
Martin, G., & Pear, J. (1998). Behavior Modification: What It Is and How to Do It (6th ed.). Prentice Hall.

← Previous · II-005Vittoz method Next · II-007 →Counterconditioning