Coach, not tester
A tester asks a question they already know the answer to, then waits to see whether the other person can produce it. A coach hands over whatever is needed and makes the right answer happen.
Everything else about home practice is detail. This is the part that decides whether the fifteen minutes helps or hurts, and it is also the part that decides what happens to your relationship over the next few years.
- Let the silence run. Count to five before you say anything. Ten is better.
- Give the sound, not the answer, and give hints in the same order every time.
- Never quiz. No questions you already know the answer to.
- Never correct in ordinary conversation.
- Praise the attempt, not the accuracy.
- Keep practice inside its slot. Outside it, you are the spouse, not the therapist.
Why this matters more than which exercise you pick
A 2026 study by Siegle and colleagues interviewed care partners of people with aphasia in depth. It found that partners slide into being an informal therapist, and that the role conflicts with being a spouse. In the study's words, partners turn mundane interactions into therapeutic exercises, making relationships feel procedural rather than relational, and directly undermining intimacy. The themes the researchers named were shapeshifting roles, loss of intimacy, relentless positivity, and resentment.
That is a warning, and it is worth taking seriously before it happens rather than afterwards. In PPA specifically, difficulty in conversation predicts caregiver burden, and the path it takes runs through the care partner's own depression.
The ASHA review of counselling and care partner training in PPA separates two different jobs a partner does: delivering treatment that a clinician designed, and making everyday conversation work. They are different hats and they need different rules. The review recommends that clinicians fade their support as partners take on delivery, and that strained relationship dynamics and burnout get addressed head-on rather than ignored.
You can absolutely run the practice. Grasso and colleagues (2017) found that word sets trained by a caregiver did as well as sets trained by a clinician in logopenic PPA, when the caregiver followed the clinician's protocol. The evidence supports you doing this. The guardrails are entirely about how.
Let the silence run
Count to five silently before you offer anything at all. Ten is better and will feel impossible at first.
The silence is not empty. That is the retrieval attempt happening, and the effort is part of what makes practice stick. Tactus Therapy puts it bluntly for clinicians: therapists are often too quick to provide help. If you are the spouse, you are faster still. You know the word, you can see them struggling, and every instinct you have is to end the discomfort.
Do not finish their sentences unless they have asked you to. Agree in advance what they want when they get stuck. AFTD calls this a filling-in-the-blank plan: do they want you to supply the word, wait, write it down, or offer choices? The answer may be different at home and in a restaurant. Ask them while asking is still easy.
One important balance. The attempt should be invited, never required and never timed. No countdowns, no stopwatch. If they tense up, skip straight to giving them the word.
Give the sound, not the answer
When the wait is over and they want help, give the smallest useful piece rather than the whole thing, and give the pieces in the same order every single time. A fixed order stops feeling like a test and starts feeling like a routine.
This hierarchy is adapted from the cueing sequences used in the Henry lab's lexical retrieval treatment and in Tactus Therapy's clinical guidance, running from least help to most.
| Step | What you do | What it sounds like |
|---|---|---|
| 0 | Wait. Say nothing. | (five to ten seconds of silence) |
| 1 | A meaning cue, or a personal fact your family wrote for that photo | "You use it to buy things." "That's the mug Sarah gave you." |
| 2 | Finish the sentence for them to complete | "I need to earn some…" |
| 3 | The first sound | "It starts with mmm…" |
| 4 | The first letter, written where they can see it | you write m _ _ _ _ |
| 5 | The whole word, said warmly, for them to repeat | "Money. Say it with me. Money." |
Three rules go with the table. They ask for the next cue; you do not push it up the ladder. If they tense at any point, jump straight to step 5. And every card ends with the word said correctly, whatever it took to get there. Record which step got them there, and move on.
Ask the SLP which order they want, because it depends on the variant. In the logopenic variant, sound and letter cues do most of the work. In the semantic variant, where word meaning itself is fading, the meaning cues matter more. Do not invent your own order, and do not invent harder exercises when things go well. Escalation is the clinician's job.
When they get it wrong, give the correct word immediately and without criticism, have them say it, and carry on. Research on naming treatment finds that informative feedback, meaning you supply the right answer, does better than feedback that only flags the error. Aim for a success rate around 80 to 90%. If you are below that, the set is too hard or too long, and that is a problem with the set, not with them.
Never quiz
A test question is any question you already know the answer to. "What's this?" "Who's that in the photo?" "What did we have for lunch?"
When researchers asked people with aphasia, family members, and speech therapists separately what they most wanted to change about conversations, all three groups independently named test questions. They also sit on the list of conversation barriers used in the Better Conversations with PPA programme, alongside interrupting and changing the subject abruptly.
Test questions feel helpful. They feel like exercise, like keeping the mind active. What they actually do is put the person on the spot, with someone they love watching to see whether they fail. Word retrieval is a physical stressor for people with aphasia, and pressure makes retrieval worse in the moment. You are making the thing you are worried about harder.
Turn questions into statements. Instead of "who's that?", say "that's Tom at his wedding." If you need an answer, make it a yes/no question or a choice between two named things. Hand the word over and keep the conversation going.
Never correct in conversation
Practice has a format, a slot, and a purpose. Ordinary conversation has none of those, and correction does not belong in it.
AFTD's guidance is to accept what was said, no matter how imperfectly, and add to it to build the conversation. Better Conversations with PPA names message-focused communication, rather than perfect interaction, as one of its core components. The point of talking is the exchange. If you got the message, the sentence worked.
So: never correct grammar. Never correct pronunciation outside the practice slot. Never say "try again" more than once. If a word came out wrong but you understood it, let it go and answer the meaning.
When you genuinely did not understand, put the failure on yourself. "I'm not following well today, can you show me another way?" rather than "you're not making sense." And do not pretend to understand when you do not. Say so, and switch channels: point, draw, write, show me the photo.
Praise the attempt, not the accuracy
"That's right!" sounds like encouragement and lands like a verdict. It tells them you were marking. Praise the effort and the session instead: "that was a good one," "nice," "that took some doing."
Count any channel as a win. A gesture, a drawing, a written fragment, a photo pulled up on their phone, pointing at the thing. Getting the message across is the goal, and every route to it deserves the same warmth. In conversation that counts as success, and in practice it gets praised too.
Keep the numbers away from them entirely. Score the session, not the person. If you are tracking which cue level worked and how many items you got through, that record is for you and the speech-language pathologist. It never appears on their side of the table, and it never gets discussed in front of them.
No buzzers, no red marks, no "you got six out of ten today." The only feedback they need is the correct word, said warmly.
Instead of this, try this
| Instead of | Try |
|---|---|
| "What's this?" | "That's your blue mug. Sarah gave you that one." |
| "Who's that in the photo?" | "That's Tom, at his wedding." |
| "Come on, you know this." | (say nothing, and count to five) |
| "No, it's a fork." | "Here it is: fork. Say it with me. Fork." |
| "Try again." | "Here it is: fork. Again… good." |
| "Say it properly." | "I got it." |
| "Do you remember what we did yesterday?" | "We walked on the beach yesterday. Was it too windy?" |
| "You're not making sense." | "I'm not following well today. Can you show me another way?" |
| "Use your words." | "Point to it, or draw it, or write the first letter." |
| Finishing their sentence for them | "Take your time. Nothing depends on this." |
| "That's right!" | "That was a good one." |
| "Let's do ten more." | "Three easy ones to finish." |
| "You're not concentrating." | "This is frustrating. Let's leave it and come back after lunch." |
Keep practice inside its box
Separate the two roles in time and in place. Practice happens at the kitchen table, at ten in the morning, with the cards or the app, for twenty minutes. Outside that slot you are their husband or wife or son, and you use supported conversation only. No drilling at dinner. No quizzing in the car.
The Communication Bridge programme at Northwestern builds this separation into its design. The app delivers the drills; the live sessions coach the couple on strategies and on building a communication alliance. The app is the tester. You are the ally. If you are using a practice tool at home, let the tool be the one that asks.
Better Conversations with PPA works the same way from the other direction. A therapist watches video of the couple's real conversations with them, and points out what helped, such as gesture, giving time, and writing things down, and what got in the way. The partner is a participant changing their own behaviour, not an instructor grading someone else's.
Spread the load. If adult children or a friend can each take a practice slot, do it. It gives variety, it lightens the week, and it keeps the marriage from turning into a clinic.
A script for twenty minutes
If it helps, use the same words every day. Predictability is doing real work here.
Opening. "Twenty minutes, then coffee. No pressure. I'm just here with the pictures."
Each card. Show the photo. Wait, and count silently. If they say it: "Yes. Fork. Nice." If they ask for help: give the next cue in order, one step only. If they tense: "Here it is: fork. Say it with me… fork. Again… good." Move on.
Closing. "Three easy ones to finish." Then: "Done. That was a good one."
Afterwards. Log the numbers later, out of sight.
Stop the moment it stops being tolerable. A five-minute session is useful information for the clinician, not a failure. The consensus review on the semantic variant says it plainly: when practice becomes stressful, it may not be desirable to continue. If they are escalating, stop the task completely and go and do something together that needs no words. The relationship comes first, every time.
Look after the coach
A weekly self-check with three questions is enough: Did practice stay inside its slot? Did I finish their sentences? Did I take my own break?
Get into a support group and stay in it. The National Aphasia Association and Northwestern's Mesulam Center both run free online groups for PPA care partners on a monthly schedule; check the current times on their pages. The AFTD HelpLine is 866-507-7222. A pilot programme of seven weekly video sessions for spouses of people with PPA reported lower depression and less relationship strain afterwards, so ask your clinic what is running near you.
If you are at the very start of this, the companion page is What to do at home after a PPA diagnosis: the first month.
Sources
- Siegle, et al. Emotional, relational, and identity shifts in aphasia caregiving: an interpretative phenomenological analysis of care partner experiences. IJLCD 2026. PMC13352106
- Counseling and Care Partner Training in Primary Progressive Aphasia. Perspectives ASHA SIG 2021. PMC9351599
- Communication difficulties and their association with caregiving burdens in aphasic dementia. PMC8682255
- Aphasia partnership training: what outcomes do people with aphasia, family members and SLTs expect? 2025. PMC11842014
- Volkmer A, et al. Randomised controlled pilot study of Better Conversations with PPA. Pilot Feasibility Stud 2023. PMC10203671 · UCL BCPPA resources
- Rogalski EJ, et al. Communication Bridge-2 primary and secondary outcomes. Alzheimer's & Dementia 2025. PMC11712820 · trial protocol PMC9190461
- Henry ML, et al. Treatment for word retrieval in semantic and logopenic variants of PPA. 2019. PMC6802912
- Grasso, et al. 2017, on caregiver-administered word training in logopenic PPA, as summarised in the care partner training literature above.
- Middleton EL, et al. Errorless, errorful, and retrieval practice for naming treatment in aphasia: a scoping review. JSLHR 2022. PMC10023178
- Suárez-González A, et al. Semantic variant PPA consensus review. 2021. PMC8699306
- Physiological stress responses to word retrieval in aphasia. 2020. pubmed 32160628
- Tactus Therapy, cueing hierarchy for word finding. tactustherapy.com
- Aphasia Institute, Supported Conversation for Adults with Aphasia. aphasia.ca
- AFTD, Communicating with Persons Living with PPA. PDF
- PPA Tele-Savvy pilot for caregivers. Innov Aging 2023. PMC10737934
- National Aphasia Association PPA care partner support group. aphasia.org · Mesulam Center support groups