Case Study · El Salvador
Symptoms fell by half after one session.
Twenty-nine people at a children’s home in El Salvador. Seven sessions over twenty-three days. Below is everything we measured — including the parts that didn’t work.
Autumn 2022 · La Casa de mi Padre and ONUVA · delivered by Tiffany Lewis, LMFT
29
participants
children, house parents, teachers
26 of 28
improved
92.9% of those scored
49%
after one session
mean symptom reduction
74%
by final session
mean reduction from baseline
Two of the twenty-nine sit outside that denominator: one recorded no symptoms at any session, and one finished above where they started. Both are shown in full below.
The Setting
Everyone in the building was carrying something.
La Casa de mi Padre is a children’s home in El Salvador. The courts place children there after abuse, neglect, or abandonment. Three houses, roughly eight children in each, ages two to seventeen — alongside house parents, teachers, and support staff at the partner school, ONUVA.
The adults were carrying their own histories while holding the children’s. That is why the study enrolled both. If regulation only reaches the children, it is working against the room.
More than half of Salvadoran young people have been exposed to significant trauma. Talk therapy, where it exists at all, asks for language, months, and money. This asked for none of those.
What Direct Neurofeedback is
A very small signal is delivered through sensors on the scalp. It is painless and usually not even noticeable. The brain detects it and self-corrects, releasing patterns it has been stuck in.
Sessions run about thirty minutes. There is nothing to say and nothing to relive — which is why it reaches people talk therapy cannot: toddlers, non-verbal children, anyone who has run out of willingness to tell the story again.
Participants rated five symptoms — anxiety, insomnia, anger, depression, and migraine or other physical complaints — from 0 to 10 before each session. The composite score is those five added together, so it runs 0 to 50.
The Results
Most of the change happened in the first week.
The composite symptom score, averaged across everyone who attended each session. The steepest drop is between session one and session two — a mean reduction of 49.4%, measured a week later rather than in the moment.
Ratings were taken before each session, so every drop shown here is what carried over between sessions, not how someone felt walking out of one.
Mean composite symptom score (0–50)
Session five goes the wrong way, and that is the schedule, not the protocol. Sessions were not evenly spaced. Sessions two, four and seven were rated the day after the previous session; three, five and six about a week after. Across the whole study, scores rose at 44% of week-gap steps and only 14% of next-day steps. The benefit was real and it was fading between visits — which is exactly what the next chart is about.
View as a table
| Session | Same 19 people | Everyone who attended | n (all) |
|---|---|---|---|
| Session 1 | 16.74 | 14.66 | 29 |
| Session 2 | 8.53 | 7.66 | 29 |
| Session 3 | 6.58 | 6.61 | 28 |
| Session 4 | 4.26 | 5.62 | 26 |
| Session 5 | 8.42 | 7.2 | 25 |
| Session 6 | 3.68 | 3.5 | 20 |
| Session 7 | 1.75 | 1.75 | 8 |
Durability
Each session held longer than the one before it.
The practitioner recorded how long each session’s effect lasted. After the first session, the benefit held about two days. By the fifth, a week. That is the answer to the question funders ask first — what happens when the sessions stop — and the answer is that people move onto a maintenance schedule rather than a permanent one.
Median days the effect held, by session
Read this as a direction, not a dose curve. You cannot observe an effect lasting longer than the gap before the next measurement, and those gaps grew as the study went on. Establishing the real durability needs a follow-up study with fixed intervals. We have not run one yet.
By Symptom
Sleep moved first and moved furthest.
Anxiety was the most common complaint at intake. Insomnia was the one that resolved most completely — from a mean of 5.05 to 0.58 out of 10. Children who could not sleep in a building where they were finally safe began sleeping through the night.
Mean rating, first session to last (0–10)
First session Final sessionSymptom-level ratings exist for 25 of the 29 participants, so these five figures rest on a smaller sample than the headline numbers. Each row shows its own n. Only participants who reported a given symptom at intake are counted in that row.
Every Participant
Nobody is averaged away.
Every person in the study, ranked by how far their symptoms fell. Twelve reached zero. One finished above where they started. We are showing you that one, because a chart where everybody wins is a chart nobody should believe.
Change from first session to last, per participant
Participants are anonymised and numbered. One further participant recorded zero symptoms at every session and cannot be expressed as a percentage change; they are excluded here and from the 92.9% figure, which is why the denominator is 28 rather than 29.
In the Room
What the numbers were actually made of.
A teacher having two to three grand mal seizures a week stopped having them after her first session. She did not mention it to the practitioner for three weeks — epilepsy carries stigma — and put the change down to being more patient in class.
A twelve-year-old boy’s chronic bedwetting stopped after one session. By session four he rated his anger at 8 out of 10. That was not a relapse. It was a boy who had been numb starting to feel things again, and it is the clearest example of why a symptom score can understate what is happening.
Children as young as eighteen months responded — fewer tantrums, better sleep, steadier mood — without a word being exchanged about any of it.
The house parents and teachers got calmer too. That is not a footnote. Regulated adults are the environment these children live inside, and it compounds.
“In every student and teacher who has come through, levels of worry, anxiety, sadness, and insomnia have decreased and happiness keeps growing.”
— Educator at ONUVA, El Salvador
From the session notes
- “felt calm wave”
- “recover from anger quickly. Was able to clean closet”
- “husband commented on her calmness”
- “no migraines for 8 days”
- “able to disconnect and rest on day off”
- “sábado y domingo con mucha energía” — Saturday and Sunday with lots of energy
- “más ánimo” — better spirits
- “didn’t feel any change during session” — the same participant reported sleeping great at the next one
Side Effects
About one session in ten had one. All of them passed.
Across 149 recorded sessions there were roughly fifteen reports of a side effect: tingling in the hands, a mild headache lasting ten to twenty minutes, a headache the following day, tiredness, or pressure at the electrode sites.
One was more than mild: a participant had a migraine and vomited after her sixth session. It did not recur, and her symptoms had fallen from 34 to 2 by the end.
One participant noted increased anxiety at night once the effect wore off — the same fading we see in the session curve, described from inside it.
Limitations
What this study cannot tell you.
- No control group. Everyone received the protocol. We cannot separate its effect from time, attention, or expectation.
- Self-reported. Symptoms were rated by participants, some of them children, in a setting where the adults around them were visibly hopeful.
- Small and short. Twenty-nine people over twenty-three days. Eight reached a seventh session.
- Uneven spacing. Sessions were scheduled around a working school, not a research design, which is why the curve saw-tooths.
- The measure has a blind spot. A symptom score understates improvement in someone whose baseline was flat because they were numb. At least one participant’s rising score was awareness returning, not deterioration.
What would settle it: a controlled study with fixed session intervals and follow-up past the protocol. We would like to help fund one.
What Happens Next
This is twenty-three days out of a much longer story.
The study you have just read sits inside an ongoing deployment in El Salvador that has delivered more than 18,000 sessions. The government is now looking at direct neurofeedback for its developing foster care and prison systems. Roughly 60,000 children were left without primary caregivers after the gang crackdown.
Restore the Last funds the technology, the training, and the ongoing support into organisations already doing the work. They bring the people. The evidence is not thin. The gap is not proof. The gap is resources.
The Full Study
Take the case study with you.
The complete study as a printable PDF — every chart on this page, plus the methodology behind the numbers.
Gets you the PDF now, plus Field Notes — one email a month from the field. No spam, unsubscribe anytime. We never sell or share your email — privacy.
Your download has started. If it didn’t, download it directly.
Something went wrong. Please try again, or email dallas@restorethelast.org.
Every figure on this page is generated directly from the practitioner’s session records. Methodology and the underlying analysis are available on request.